Forensic Science International

Forensic Science International is an international journal publishing original contributions in the many different scientific disciplines pertaining to the forensic sciences. Fields include forensic pathology and histochemistry, chemistry, biochemistry and toxicology (including drugs, alcohol, etc.), biology (including the identification of hairs and fibres), serology, odontology, psychiatry, anthropology, the physical sciences, firearms, and document examination, as well as investigations of value to public health in its broadest sense, and the important marginal area where science and medicine interact with the law. Review Articles and Preliminary Communications (where brief accounts of important new work may be announced with less delay than is inevitable with major papers) may be accepted after correspondence with the appropriate Editor. Case Reports will be accepted only if they contain some important new information for the readers. Submission of Articles: Manuscripts prepared in accordance with Instructions to Authors should be sent to the Editor-in-Chief or one of the Associate Editors according to their area of expertise as listed below. If there is any doubt, a preliminary letter, telefax or telephone enquiry should be made to the Editor-in-Chief or his Assistant.

EDITOR-IN-CHIEF P. Saukko – (for: Experimental Forensic Pathology, Traffic Medicine and subjects not listed elsewhere) Department of Forensic Medicine, University of Turku, SF-20520 Turku, Finland Tel.: (+358) 2 3337543; Fax: (+358) 2 3337600; E-mail: [email protected]

SECRETARY TO THE EDITOR-IN-CHIEF A. Saarenpää – Address as for P. Saukko. Tel: (+358) 2 3337438; Fax: (+358) 2 3337600; E-Mail: [email protected]

ASSOCIATE EDITORS A. Carracedo – ( for: Forensic Genetics) Instituto de Medicina Legal, Facultad de Medicina, 15705 Santiago de Compostela, Galicia, Spain Fax: (+34) 981 580336 O.H. Drummer – ( for: Toxicology) Department of Forensic Medicine, Victorian Institute of Forensic Medicine 57–83 Kavanagh Street, Southbank 3006, Victoria, Australia Tel.: (+61)-3-9684-4334; Fax: (+61)-3-9682-7353

P. Margot – (for: Questioned Documents, with the assistance of A. Khanmy and W. Mazzela; and for Physical Science: ballistics, tool marks, contact traces, drugs analysis, fingerprints and identification, etc.) Ecole de Sciences Criminelles (ESC), UNIL-BCH, CH-1015 Lausanne, Switzerland Fax: (+41) 21 692 4605 G. Willems – ( for: Odontology) Katholieke Universiteit Leuven, School of Dentistry, Oral Pathology and Maxillo-Facial Surgery, Departments of Orthodontics and Forensic Odontology, Kapucijnenvoer 7, B-3000 Leuven, Belgium Tel: +32 16 33.24.59; Fax: +32 16 33.24.35

Cristina Cattaneo – ( Anthropology and Osteology) Instituto de Medicina Legal, Universit`a degli Studi Via Mangiagalli 37, 20133 Milano, Italy Tel: 0039 2 5031 5678; Fax: 0039 2 5031 5724

EDITORIAL BOARD A. Amorim (Porto, Portugal) J. Buckleton (Auckland, New Zealand) B. Budowle (Quantico, VA, USA) J. Clement (Melbourne, Australia) S.D. Cohle (Grand Rapids, MI, USA) S. Cordner (South Melbourne, Australia) P. Dickens (Buxton, UK) M. Durigon (Garches, France) A. Eriksson (Umeä, Sweden) J.A.J. Ferris (Auckland, New Zealand) M.C. Fishbein (Encino, USA) P. Gill (Solihull, UK) C. Henssge (Essen, Germany) M.A. Huestis (Baltimore, MD, USA)

A.W. Jones (Linköping, Sweden) H. Kalimo (Helsinki, Finland) B. Kneubuehl (Thun, Switzerland) S. Leadbeatter (Cardiff, UK) P.J. Lincoln (Surrey, UK) A. Luna Maldonado (Espinardo (Murcia), Spain) B. Madea (Bonn, Germany) N. Morling (Copenhagen, Denmark) B. Olaisen (Oslo, Norway) V. Pascali (Rome, Italy) S. Pollak (Freiburg i. Br., Germany) M.S. Pollanen (Toronto, Canada) D. Pounder (Dundee, UK) O. Prokop (Berlin, Germany)

K. Püschel (Hamburg, Germany) G. Quatrehomme (Nice, France) J. Robertson (Canberra, Australia) P.M. Schneider (Cologne, Germany) S. Seta (Tokyo, Japan) J. Simonsen (Copenhagen, Denmark) P. Sótonyi (Budapest, Hungary) M. Steyn (Pretoria, South Africa) F. Tagliaro (Verona, Italy) T. Takatori (Chiba, Japan) S. Tsunenari (Kumamoto, Japan) D.N. Vieira (Coimbra, Portugal) X. Xu (Shantou, People’s Republic of China) J. Zhu (Guangzhou, People’s Republic of China)

Forensic Science International 166 (2007) 85–90 www.elsevier.com/locate/forsciint

Lethal paradoxical cerebral vein thrombosis due to suspicious anticoagulant rodenticide intoxication with chlorophacinone F. Papin a, F. Clarot b,*, C. Vicomte b, J.M. Gaulier c, C. Daubin d, F. Chapon e, E. Vaz b, B. Proust b a Forensic Department, Caen University Hospital, Caen, France Medical Forensic Institute, Rouen University Hospital, Charles Nicolle, Rouen, France c Pharmacokinetic and Toxicology Laboratory, Limoges University Hospital, Limoges, France d Intensive Care Unit, Caen University Hospital, Caen, France e Neuropathology Department, Caen University Hospital, Caen, France b

Received 9 February 2006; received in revised form 4 April 2006; accepted 9 April 2006 Available online 23 May 2006

Abstract Superwarfarin exposure is a growing health problem, described in many countries. The authors report a case of suspicious chlorophacinone poisoning with a problematic diagnosis. They review the literature and discuss particularities of anticoagulant rodenticide intoxication, as well as the apparent contradiction between anticoagulant intoxication and lethal thrombosis. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Rodenticide; Chlorophacinone; Poisoning; Anticoagulant; Cerebral vein thrombosis

1. Introduction Rodenticide intoxication is rare, as these products are no longer used as rodenticide due to its hazardous effects on humans. Nevertheless, these products may still be found in certain garden sheds, and could have a ‘‘criminal’’ use. The authors report a case of suspicious chlorophacinone poisoning with a problematic diagnosis. Particularities of anticoagulant rodenticide intoxication are discussed; usefulness of blood analysis in suspected poisoning or intoxication is underlined. The authors also discuss the physiopathological characteristics of their case, as well as the apparent contradiction between anticoagulant intoxication and lethal thrombosis. 2. Case report A 34-year old woman, farm worker, with no particular previous medical history or medication – with the exception of * Correspondence to: Institut de Me´decine Le´gale, CHU Rouen, Charles Nicolle, 76031 Rouen Cedex, France. Tel.: +33 232888284; fax: +33 232888367. E-mail address: [email protected] (F. Clarot). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.003

an oral oestroprogestative contraception – presented in the emergency department with a massive hematuria. On admission, the patient was apyretic and physical examination confirmed macroscopic hematuria and had abdominal pain in the right hypochondral region, which initially occurred 3 days earlier. No particular violence related lesion was observed. The patient was initially treated by analgesics and intravenous NSAIDs. Blood laboratory tests revealed a hyperleucocytosis (12,500/ 3 mm) which resulted in antibiotic treatment and hospitalization for suspected infected renal colitis. The following day, abdominal ultrasonography was performed and revealed pyelo-calicis hyperechogenicity with no dilatation. Moreover, a slight fluid collection was observed medially to the right kidney, and a mobile echogenic residue was also found in the bladder. The patient continued to have pain and hematuria, and at day 4 she suddenly presented convulsive loss of consciousness. She was transferred to the intensive care unit in a comatose state (Glasgow scale 4) and was then sedated, intubated, and ventilated. Initial neurological examination showed hypotonic coma, a non-reactive left mydriasis and rapidly bilateral areactive mydriasis. She was administered 100 mg of mannitol and a CT

86

F. Papin et al. / Forensic Science International 166 (2007) 85–90

Fig. 1. Initial CT scan show a large left hemisphere haemorrhagic cerebral infarct, a diffuse oedema, a subarachnoid haemorrhage, and a right shifting of the midline structure with a left ventricule disappearance.

scan was performed with no contrast agent (Fig. 1). It revealed an extended left hemisphere haemorrhagic cerebral infarct, associated with diffuse oedema and midline structure shifted to the right. Biological screening revealed a slight anemia (10.3 g/ dl), an increased hyperleucocytosis (19,490/3 mm) and a normal platelet count. Coagulation study showed an unexplained very low prothrombin time (10%) and especially very low levels of Vitamin K dependant factors (II, VII, IX, and X). However, antithrombin III and factor V were normal. Cerebral angiography, performed to assess the brain perfusion state, revealed a thrombosis of the superior longitudinal sinus (SLS) (Fig. 2). Despite intravenous Vitamin K injection and adapted resuscitation, our patient died in an irreversible comatose state, at day 4.

Further subsequent toxicological analysis performed on a serum sample collected during hospitalization, the 4th day before the death, revealed a high level of chlorophacinone: 25.9 mg/L. At autopsy, performed 4 days after death, we found a slight nail and lip cyanosis, pulmonary asphyxia lesion, and a trachea oedema. Moreover, autopsy revealed diffuse haemorrhagic signs (i.e. multiple ecchymosis, visceral haemorrhages, pleural and peritoneal blood collection, diffuse subarachnoid haemorrhage, and renal intracavity haemorrhage). The biological samples collected were sent to the laboratory for forensic toxicological analysis. Macroscopic and histologic examination confirmed a bilateral and diffuse alveolar pulmonary oedema, a multivisceral congestion, and demonstrated a concentric myocardial hypertrophy. Neuropathological examination confirmed SLS thrombosis. It also showed a left frontal region haemorrhagic infarct lesion, a diffuse oedema, and herniation of the fifth temporal circumvolution (Fig. 3). Police investigation was not able to assess the origin of the intoxication, which was not considered as criminal, but accidental or suicidal. 3. Materials and methods Venous blood sample was collected during hospitalization, the fourth day before the death. Initially, blood was taken to perform coagulation tests but extensive toxicological screening (including a chlorophacinone assay) was subsequently performed in a serum sample, due to the diagnostic problems, using high-performance liquid chromatography coupled with diode-array detection (HPLC– DAD). A second toxicological investigations set was performed in a forensic context 4 days after death, on autopsy biological samples (peripheric blood, urine, pleural effusion, and gastric contents). No visceral, particularly liver, analysis was performed.

Fig. 2. Cerebral angiography (via left vertebral artery) shows lack of flow of the superior longitudinal sinus and a capilar ‘‘marshy’’ stasis (left image). Lack of SLS opacification is confirmed by left lateral sinus venography (right image).

F. Papin et al. / Forensic Science International 166 (2007) 85–90

87

Fig. 3. Macroscopic brain examination confirmed a left frontal haemorrhagic infarct lesion, extended to the caudate nucleus head and the corpus callosum. Examination showed as well a herniation of the 5th temporal circumvolution.

Chlorophacinone assays in serum were performed using a previously published method [1]. This analytical procedure was also applied for chlorophacinone determinations in forensic samples (i.e. blood, urine, and gastric contents) according to the results of a previous analytical validation step in such biological matrix. Briefly, this method, routinely applied for the simultaneous identification and quantitation of 13 hydroxycoumarin and indandione anticoagulant drugs [2] and rodenticides, used a reversed-phase liquid chromatography with diode-array detection technique. Extraction step consisted of an acidic and alkaline liquid–liquid double extraction with diethylether– ether acetate (50:50, v/v). High-performance liquid chromatography was performed using gradient elution with an acetonitrile and phosphate buffer on a Nucleosil C18, 5 mm particle size (150 mm  4.6 mm i.d.) column. Detection and quantitation limits for chlorophacinone were 20 and 50 mg/L, respectively. The standard calibration curve was linear from 50 to 5000 mg/L; within-run precision coefficient of variation (CV) was less than 10%, and between-run precision CV was less than 20%. Table 1 Chlorophacinone levels

Ante mortem samples Post mortem samples

Blood (mg/L)

Urine Urine (mg/L) (mg)

Gastric contents (mg/L)

Gastric contents (mg)

25.9

–

–

–

–

6.8

0.102

4.8

0.192

9.4

4. Results Toxicological screening revealed ante and post mortem high blood concentration of chlorophacinone (ante mortem toxicological screening was performed on venous peripheral blood; post mortem blood was obtain from subclavian artery). Citalopram and desmethyl diazepam were also discovered, but at therapeutic levels. Post mortem toxicological analyses were also performed in urine (15 mL) and gastric contents (40 mL); results are shown in Table 1. 5. Discussion Rodenticides are the name given to any of the group of toxic substances that are used to kill rodents. Rodenticides are a group of compounds that exhibit markedly different toxicities to humans and rodents. The varieties of rodenticides used over the years are numerous, leading to the popular expression, ‘‘to build a better mousetrap’’. Adults who ingest these substances are most likely individuals attempting suicide; however, poisoning homicides may occur with these agents due to their ready availability. Superwarfarin exposure is a growing health problem [3], described in many countries. In 2002, in U.S.A., according to the Toxic Exposure Surveillance System (TESS) of the American Association of Poison Control Centers (AAPCC), 19,674 human exposures to rodenticides have been reported. According to the 2002 TESS data, anticoagulant rodenticides were associated with 16,822 of rodenticide exposures, but only two lethal intoxications were observed [2]. These cases are

88

F. Papin et al. / Forensic Science International 166 (2007) 85–90

always reported as accidental exposition, suicide attempts or Mu¨nchausen syndrome [4–8]. At the turn of the century, rodenticides were composed heavy metals such as arsenic, thallium and phosphorus along with red squill and strychnine. This changed in the 1940s as investigators discovered that warfarin could be transformed in bishydroxycoumarin when fungi in moldy sweet clover oxidize coumarin to 4-hydroxycoumarin. Warfarin was quickly adopted as the major rodenticide and in 1940, bishydroxycoumarin was synthesized and used clinically 1 year later as an oral anticoagulant under the American trade name dicumarol. However, rodents resistance to warfarin became prevalent in the 1960s via autosomal dominant gene transmittance [9]. Novel compounds were synthesized to combat rodent resistance, thereby creating a new class of anticoagulants— the superwarfarins [10]. The term superwarfarin refers to a group of compounds, second-generation anticoagulants, which are extremely longacting. These fat-soluble anticoagulants are colorless, tasteless, and odorless compounds [11]. Superwarfarins, as warfarin, inhibit hepatic synthesis of the Vitamin K-dependent coagulation factors II, VII, IX, and X and the anticoagulant proteins C and S. Chlorophacinone stops the synthesis of the active form of Vitamin K1 via inhibition of Vitamin K1–3 epoxide reductase, which blocks coagulation factor synthesis (II, VII, IX, and X) [2,9,12–16]. Superwarfarins are metabolized by hepatic cytochrome P450 isoenzymes to hydroxylated metabolites. It is uncertain that whole metabolites are inactive [9]. Chlorphacinone is an indandione-derivative with a prolonged effect of the superwarfarins family [11]. This anticoagulant is approximately 100 times more potent than warfarin on a molar basis [9]. The half-life of superwarfarin varies from 16 to 69 days compared with 37 h for warfarin [16,17]. The most human toxic form is an oily base (concentration of 2.5 g/ L), which is found in numerous commercial products worldwide. However, this form is no longer sold in France (since 2000) as it has been considered a health hazard. Each Vitamin K-dependent factor differs in its degradation half-life; factor II requires 60 h, factor VII requires 4–6 h, factor IX requires 24 h, and factor X requires 48–72 h. The half-lives of proteins C and S are approximately 8 and 30 h, respectively. As a result, because antivitamin K reduces first the activity of anticoagulant proteins C and S, a hypercoagulable state may be initially induced. Rapid loss of protein C temporarily shifts the balance in favour of clotting until sufficient time has passed for antivitamin K to decrease the activity of coagulant factors [18–21]. Chlorophacinone poisoning induces prolonged prothrombin time (PT), elevated international normalized ratio (INR), extended activated prothromboplasmin time, and decreased Vitamin K-dependent factors levels. Bleeding is the most common clinical feature and may occur from any mucosal site or organ [22–25]. The first haemorrhagic signs usually occur 3– 7 days after intake, depending on the dose ingested, and the substance half-life (from 6 to 23 days), when the body’s

reserves of prothrombin have diminished [2,4,9,12–14,26–30]. Table 2 showed 16 cases of chlorophacinone intoxication described in the literature. 5.1. Our case raises multiple problems First, concerning the origin of the intoxication, which was not (and will probably never been) assessed. Because our patient, and her family, were farm workers, they consequently had access to concentrated rodenticide for professional use. Regarding the elevated blood level of chlorophacinone, it is uncertain that intoxication involved granules, because this volume would have been too bulky to ingest. Only the oil form is known to be sufficiently concentrated to be hazardous, in small quantities, able to be ingested ‘‘accidentally’’. However, an accidental ingestion of oily chlorophacinone would suppose package reconditioning, or a severe neurological state impairment. Suicidal intoxication was considered but our patient had no previous suicidal history, nor psychiatric symptoms. Criminal poisoning was also considered, but police investigations found no arguments in favour of this hypothesis. Second, the coexistence of a haemorrhagic syndrome and an anticoagulant intoxication was initially disturbing. However, the first hypothesis considered was that SLS thrombosis occurred initially and subsequently induced neuropsychiatric impairment. Accidental or suicidal ingestion would have been consequent to these alterations. Nevertheless, our patient’s husband did not described major or sufficient behavioural abnormalities in favour of this hypothesis. In fact, the review of the literature regarding warfarin has explained this apparent contradiction [31–33]. Certain studies involved warfarin levels, monitored by measuring the prothrombin time, which responds to reductions in levels of three Vitamin K-dependent clotting factors (factors II, VII, and X). It has been demonstrated that during the first 48 h of treatment, the anticoagulant effect of warfarin is caused mainly by a reduction in the activity of factor VII, which has a half-life of 6 h. In contrast, the antithrombotic effect of warfarin (which is thought to be caused primarily by a reduction in the activity of factor II) is delayed for as long as 60 h. Therefore, during the first 48 h of therapy, the anticoagulant and antithrombotic effects of warfarin may be unrelated. In addition, because the half-life of the Vitamin K-dependent anticoagulant protein, protein C, is similar to that of factor VII, the early anticoagulant effect of warfarin (which results from reduction of factor VII) could be counteracted by a procoagulant effect (which results from reduction of protein C). Moreover, it has been also demonstrated that greater dose of warfarin was associated with a significantly more rapid decrease in protein C activity (which decreased before levels of factors X and II were substantially reduced). Therefore, the combination of markedly reduced protein C levels and near-normal levels of factors II and X over the first 2 days of warfarin therapy could produce an initial hypercoagulable state. Obviously, it is not possible to study chlorophacinone effects on humans, but it is probably scientifically possible to extrapolate warfarin data to superwarfarins.

F. Papin et al. / Forensic Science International 166 (2007) 85–90

89

Table 2 Review of the literature (when chlorophacinone blood level or absoption is reported) Reference

# Case

Sex (age (years))

Intoxication reason

Death (Yes/No)

Absorption mode, dose ingested

Symptoms (appearance delay)

Biological tests (appearance delay)

Chlorophacinone blood level, half-life (Hl)

Cutaneous mucous haemorrhage (day 10) None

PT = 0%, APT = 65/32 (day 10) INR = 1.2 (hour 18)

Unknown Unknown

Haematomas, ecchymosis (day 37) None Multiple cutaneous mucous haemorrhage (day 16) Haematuria (day 15)

PT = 92% (hour 1.5)

Unknown

PT = 43% (hour 18) PT = 75% (hour 1.5); hypocoagulability (day 2) PT = 53% (a few hours later) PT = 38% (day 2) PT = 18% (day 2) PT = 65% (day 16)

Unknown Unknown

Dusein et al. [26] Murdoch et al. [30]

1

M (28)

Suicide

N

Oral, Unknown

2

F (37)

Suicide

N

Oral, 625 mg

Chataigner et al. [4]

3

F (79)

Suicide

N

Oral, 375 mg

4 5

M (45) M (18)

Suicide Suicide

N N

Oral, 500 mg Oral, 625 mg

6

M (21)

Suicide

N

Oral, 750 mg

7 8 9

M (54) M (27) M (72)

Suicide Suicide Suicide

N N N

Oral, 750 a` 1500 mg Oral, Unknown Oral, Unknown

10

M (52)

Suicide

N

Oral, 500–1200 mg

11

M (61)

Suicide

N

Oral, 500 mg

12

F (20) ans

Suicide

N

Oral, 250 mg

13

F (60) ans

Suspected suicide

N

Unknown, unknown

14

M (23) ans

Suspected suicide

N

Unknown, unknown

Arditti et al. [12]

15

F (27) ans

Suicide

N

Oral, 625 mg

Lagrange et al. [14]

16

M (33) ans

Suicide

N

Oral, 1875 mg

Burucoa et al. [9]

None None Microscopical haematuria (day 2) Haematuria, haematomas (day 2) Haematuria (day 19); gums bleeding (day 21)

Unknown Unknown Unknown Unknown

PT = 100% (a few hours later) PT < 10% (day 21)

Unknown Unknown

Lumbar pain, macroscopical haematuria (day 7) Macroscopical haematuria, vaginal bleeding, gum bleeding, thigh ecchymosis Abdominal pain, macroscopical haematuria, gum bleeding

Hypocoagulability (day 7)

Day 12: 2 mg/L, Hl: 6.5 days

Hypocoagulability

12.9 mg/L, Hl: 22.8 days

Hypocoagulability

1.2 mg/L, Hl: 11 days

Asthenia, inhalation pneumopathy (hour 80) Nausea (hour 8)

PT < 10%, APT = 44/32 (hour 80)

Hour 80: 43 mg/L, Hl: 7.6 days

PT normal (hour 8)

Hour 8: 27.6 mg/L

INR: international normalized ratio; PT: prothrombin time; APT: anti-prothrombin time.

6. Conclusion In our case, the origin of the intoxication remains unclear, and will probably never elucidated. Physicians and pathologists should bear in mind that any patient presenting with prolonged and/or unexplained hypocoagulability could have ingested superwarfarins, whether or not voluntarily. Moreover, pathologists and forensic practitioner should also keep in mind that thrombosis could paradoxically be related to anticoagulant intoxication, which is more uncommon. Acknowledgment The authors thank Richard Medeiros, Rouen University Hospital, editor, for his valuable advice in editing the manuscript.

References [1] H. Lofti, M.F. Dreyfuss, P. Marquet, J. Debord, L. Merle, G. Lachatre, A screening procedure for the determination of 13 oral anticoagulants and rodenticides, J. Anal. Toxicol. 20 (1996) 93–100. [2] W.A. Watson, et al., 2003 Annual report of the American association of poison control centers toxic exposure surveillance system, Am. J. Emerg. Med. 22 (2004) 335–404. [3] A.E. Rauch, R. Weininger, D. Pasquale, P.T. Burkart, H.G. Dunn, C. Weissman, E. Rydzak, Superwarfarin poisoning: a significant public health problem, J. Community Health 19 (1994) 55–65. [4] D. Chataigner, R. Garnier, J. Elmalem, M.L. Efthymiou, Prolonged hypocoagulability following the ingestion of anticoagulant raticides, Ann. Med. Interne (Paris) 139 (1989) 537–541. [5] J. Babcock, K. Hartman, A. Pedersen, Rodenticide-induced coagulopathy in a young child. A case of Munchausen syndrome by proxy, Am. J. Pediatr. Hematol. Oncol. 15 (1993) 126–130. [6] G. Shepherd, W. Klein-Schwartz, B.D. Anderson, Acute unintentional pediatric brodifacoum ingestions, Pediatr. Emerg. Care 18 (2002) 174–178.

90

F. Papin et al. / Forensic Science International 166 (2007) 85–90

[7] D. Kanabar, G. Volans, Accidental superwarfarin poisoning in children— less treatment is better, Lancet 360 (2002) 963. [8] M. Ingels, C. Lai, W. Tai, B.H. Manning, C. Rangan, S.R. Williams, A.S. Manoguerra, T. Albertson, R.F. Clark, A prospective study of acute, unintentional, pediatric superwarfarin ingestions managed without decontamination, Ann. Emerg. Med. 40 (2002) 73–78. [9] C. Burucoa, P. Mura, R. Robert, C. Boinot, S. Bouquet, A. Piriou, Chlorophacinone intoxication. A biological and toxicological study, J. Toxicol. Clin. Toxicol. 27 (1989) 79–89. [10] C.R. Routh, D.A. Triplett, M.J. Murphy, L.J. Felice, J.A. Sadowski, E.G.T. Bovill, Superwarfarin ingestion and detection, Am. J. Hematol. 36 (1991) 50–54. [11] C.H. Hui, A. Lie, C.K. Lam, C. Bourke, ‘‘Superwarfarin’’ poisoning leading to prolonged coagulopathy, Forensic Sci. Int. 78 (1996) 13–18. [12] J. Arditti, M. Hayek, J.H. Bourdon, G. Lachaˆtre, D. Masset, J.M. David, J. Jouglard, Self poisoning by chlorophacinone. Clinical and analytical data, Toxicorama 9 (1997) 266–270. [13] R.S. Hoffman, Anticoagulants, in: L.R. Goldfrank (Ed.), Goldfrank’s Toxicologic Emergencies, sixth ed., Appleton and Lange, Stamford, CT, 1998, pp. 703–721. [14] F. Lagrange, A.G. Corniot, K. Titier, R. Bedry, F. Pehourcq, Toxicological management of chlorophacinone poisoning, Acta Clin. Belg. (Suppl. 1) (1999) 13–16. [15] J.J. Vogel, P. De Moerloose, C.A. Bouvier, J. Gaspoz, P. Riant, Prolonged anticoagulation following chlorophacinone poisoning, Schweiz Med. Wochenschr. 118 (1988) 1915–1917 (Abstract). [16] A.M. Breckenridge, S. Chlolerton, J.A.D. Hart, B.K. Park, A.K. Scott, A study of the relationship between the pharmacokinetics and the pharmacodynamics of the 4-hydroxycoumarin anticoagulants warfarin, difenacoum and brodifacoum in the rabbit, Br. J. Pharmacol. 84 (1985) 81–91. [17] H. Thijssen, L.G. Baars, Tissue distribution of selective warfarin binding sites in the rat, Biochem. Pharmacol. 42 (1991) 2181–2186. [18] J. Hellemans, M. Vorlat, M. Verstraete, Survival time of prothrombin and factors VII, IX, X after complete synthesis blocking doses of coumarin derivatives, Br. J. Haematol. 9 (1963) 506–512. [19] H.R. Roberts, E. Lechler, W.P. Webster, G.D. Penick, Survival of transfused factor X in patients with Stuart disease, Thromb. Diath. Haemorrh. 13 (1965) 305–309.

[20] S. Vigano, P.M. Mannucci, S. Solinas, Decrease in protein C antigen and formation of an abnormal protein soon after starting oral anticoagulant therapy, Br. J. Haematol. 157 (1984) 213–220. [21] K. Okajima, S. Koga, M. Kaji, M. Inoue, T. Nakagaki, A. Funatsu, H. Okabe, K. Takatsuki, N. Aoki, Effect of protein C and activated protein C on coagulation and fibrinolysis in normal human subjects, Thromb. Haemost. 63 (1990) 48–53. [22] J.A. Kruse, R.W. Carlson, Fatal rodenticide poisoning with brodifacoum, Ann. Emerg. Med. 21 (1992) 331–336. [23] N. Nighoghossian, J.H. Ruel, P. Ffrench, P. Trouillas, Cervicodorsal subdural hematoma caused by coumarinic rodenticide poisoning, Rev. Neurol. 146 (1990) 221–223. [24] B.W. Morgan, C. Tomaszewski, I. Rotker, Spontaneous hemoperitoneum from brodifacoum overdose, Am. J. Emerg. Med. 14 (1996) 656–659. [25] J.D. Chua, W.R. Friedenberg, Superwarfarin poisoning, Arch. Intern. Med. 158 (1998) 1929–1932. [26] P. Dusein, G. Manigand, J. Taillandier, Severe, prolonged hypothrombinemia following poisoning by chlorophacinone, Presse Med. 13 (1984) 1845. [27] T.L. Abell, K.S. Merigian, J.M. Lee, J.M. Holbert, J.W. McCall, Cutaneous exposure to warfarin-like anticoagulant causing an intracerebral haemorrhage: a case report, J. Toxicol. Clin. Toxicol. 32 (1994) 69–73. [28] A.F. Pelfrene, Synthetic organic rodenticides, in: W.J. Hayes, Jr., E.R. Laws, Jr. (Eds.), Handbook of Pesticide Toxicology, Academic Press, San Diego, 1991, p. 1271. [29] R.C. Baselt, Disposition of Toxic Drugs and Chemicals in Man, Chemical Toxicology institute, Foster City, CA, 1994, p. 777. [30] D.A. Murdoch, Prolonged anticoagulation in chlorophacinone poisoning, Lancet 1 (1983) 355–356. [31] J. Hirsh, J.E. Dalen, D. Deykin, L. Poller, H. Bussey, Oral anticoagulants. Mechanism of action, clinical effectiveness, and optimal therapeutic range, Chest 108 (1995) 231S–246S. [32] A. Zivelin, L.V. Rao, S.I. Rapaport, Mechanism of the anticoagulant effect of warfarin as evaluated in rabbits by selective depression of individual procoagulant Vitamin K-dependent clotting factors, J. Clin. Invest. 92 (1993) 131–140. [33] L. Harrison, M. Johnston, P. Massicotte, M. Crowther, K. Moffat, J. Hirsh, Comparison of 5-mg and 10-mg loading doses in initiation of warfarin therapy, Ann. Int. Med. 126 (1997) 133–136.

Forensic Science International 166 (2007) 91–101 www.elsevier.com/locate/forsciint

Detection of gunpowder stabilizers with ion mobility spectrometry C. West *, G. Baron, J.-J. Minet De´partement des Explosifs et Incendies, Laboratoire Central de la Pre´fecture de Police, 39bis rue de Dantzig, 75015 Paris, France Received 11 November 2005; received in revised form 5 April 2006; accepted 7 April 2006 Available online 7 July 2006

Abstract This study is the first reported ion mobility detection of ethyl centralite and diphenylamine (DPA) smokeless gunpowder stabilizers, together with the nitroso and nitro derivatives of diphenylamine. First, the applicability of the ion mobility spectrometry (IMS) for the substances of interest was determined. The existence of numerous peaks, both in positive and negative modes, clearly demonstrates the success of these experiments. All mono and di-nitro derivatives of DPA tested were detected with this method. Unfortunately, many of the ions generated were not accurately identified. However, reduced mobility constants representative of each ion generated under defined operating conditions could be used for purpose of compound identification. The method was then successfully tested on real gunpowder samples. By the use of IMS, we managed to establish a rapid, simple and sensitive screening method for the detection and identification of smokeless gunpowder organic components. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Diphenylamine; Ethylcentralite; Smokeless gunpowder; Ion mobility spectrometry (IMS)

1. Introduction Smokeless powders are commonly used in modern ammunition. Additionally, they are frequently used in the construction of improvised explosive devices (IEDs) related to criminal and terrorist acts. Therefore, in the latter case and in firearm discharge cases, the identification of some components that can associate residue samples with unfired gunpowder provides valuable evidence to the forensic scientist. When inorganic gunshot residues have not been recovered or their characteristics are non-specific, the analysis of organic residues is required to provide complementary information. The main component of smokeless gunpowder is nitrocellulose (NC). In single-based propellants, it is the only energetic material in the composition, while in double-based powders, nitroglycerin (NG) is also present. In triple-based gunpowders, other explosive components can be found, nitroguanidine being the most frequent. However, NC and NG have been regarded as being a lack of conclusive evidence [1] as NC is widely used in

* Corresponding author. E-mail addresses: [email protected] (C. West), [email protected] (J.J. Minet). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.004

varnishes, celluloid films and pharmaceutical industry, while NG also occurs in pharmaceutical preparations [2]. Organic additives, such as dinitrotoluene (DNT) isomers— used as a flash suppresser, dibutylphthalate—used as a plasticizer, methyl and ethyl centralite (EC) and diphenylamine (DPA)—used as stabilizers are also present in gunpowder composition. EC is regarded as a characteristic material in gunpowder but it is not present in all compositions. DPA is also commonly used in the perfumery, in the food industry and as antioxidant in the rubber and elastomer industry [3], thus, the single detection of DPA is not a diagnostic of gunpowder presence. It is well known that DPA stabilizes the energetic composition by binding nitrous oxide gases originating from NC decomposition, and converting them into stable compounds. The main reaction products of nitrous oxide gases and DPA (see Fig. 1) are N-nitroso-diphenylamine (N-NODPA), 2-nitrodiphenylamine (2-NDPA) and 4-nitrodiphenylamine (4-NDPA). N-NODPA is believed to be the primary intermediate before the nitro derivatives are formed by a Fischer–Hepp rearrangement – leading to 2-nitroso-diphenylamine (2-NODPA) and 4-nitrosodiphenylamine (4-NODPA) – and an oxidation step [4,5]. These nitroso and nitro derivatives also act as stabilizers [5,6], and

92

C. West et al. / Forensic Science International 166 (2007) 91–101

Fig. 1. Dominating reaction routes of the initial DPA degradation in an aging gunpowder [4–6]. (1) DPA, (2) N-NODPA, (3) 2-NDPA, (4) 4-NODPA, (5) 4-NDPA, (6) 2-NDPA, (7) 2,4-NDPA, (8) 2,40 -NDPA.

further react with nitrous oxide gases to form highly nitrated derivatives of DPA, such as N-nitroso-2-nitro-diphenylamine (NNO-2-NDPA), 2,4-dinitrodiphenylamine (2,4-NDPA) and 2,40 dinitrodiphenylamine (2,40 -NDPA). The last compound to be formed would be the 2,20 ,4,40 ,6,60 -hexanitrodiphenylamine, that would later face decomposition into picric acid. The other origins of DPA, possibly causing environmental contamination in casework samples, can be minimized if the nitrated derivatives of DPA can also be identified in the sample, as those are unique to smokeless gunpowder. Therefore, if both energetic compounds such as NC and NG, and organic additives such as EC or DPA and some of its nitrated derivatives can be identified simultaneously in casework samples, the presence of gunpowder is ascertained. Because the amount of stabilizer initially introduced in the composition is very small (about 2%), a highly sensitive analytical method is required. The use of gas chromatography (GC) for the analysis of DPA derivatives is limited by the low volatility of the highly nitrated derivatives. Indeed, tri- and tetra-nitrodiphenylamine analogs need very high operating temperatures (up to 320 8C) [4]. GC with thermal energy analysis detection (GC–TEA) is generally prefered to GC coupled to mass spectrometry (GC–MS) as the latter is reported not to be sensitive enough for real-life samples [7]. Furthermore, N-NODPA and other nitroso derivatives of DPA are generally reported to degrade in the injection port in GC. The failure of GC to elute N-NODPA intact severely restricts its usefulness in the analysis of DPA derivatives, as N-

NODPA is the primary intermediate formed and is therefore present in high proportions. Normal phase and reversed phase high-performance liquid chromatography (HPLC) [5,8] are both commonly applied, with varied detectors: either with UV–vis spectrophotometry [9], amperometric detection [5,10], fluorimetric detection [11] or more recently with mass spectrometry [12]. HPLC leads to satisfying separations of the major nitro and nitroso derivatives. Equally good separations were reported with supercritical fluid chromatography [4,13] and micellar electrokinetic chromatography [8]. A tandem MS method was recently reported [14]. In the forensic laboratory, casework samples are often complex mixtures issued from the solvation of a complex and dirty matrix. In this case, thin layer chromatography (TLC) is the most appropriate technique. In our laboratory, we traditionally use TLC, followed by an oxidation step, for the detection of stabilizers. This method suffers from poor sensitivity and is very time consuming, but it is the most appropriate for the analysis of dirty samples. Additionally, in case studies, methods of identification based on noncorrelated separation mechanisms are needed. A method that would provide orthogonal identifications to the classical chromatographic methods is required. Thus, the combination of two such methods would enhance the identification power and the analytical reliability. Ion mobility spectrometry (IMS) is commonly used by forensic scientists to identify the major explosives, narcotics and chemical warfare agents. Basically, IMS refers to the

C. West et al. / Forensic Science International 166 (2007) 91–101

93

Fig. 2. Schematic representation of an IMS detector cell, adapted from ref. [21].

detection and characterization of chemical substances from their gas-phase ion mobilities at atmospheric pressure under an applied electrostatic field. The principles of IMS have been thoroughly described elsewhere [15], therefore, only a brief description of the principles of operation will be presented here. Because of its low cost of operation, fast response time (less than 10 s) and reduced sample preparation, IMS is particularly interesting as a screening technique. Amines and amides generally provide intense response with positive polarity in IMS because of high proton affinities with comparatively long-lived product ions [16]. EC is thus expected to produce positive ions. Although DPA response in positive ion mode has been reported [17,18], response of IMS towards its nitro derivatives has been unexplored. Nitrated compounds such as explosives have high electron affinities and are generally best observed as negative ions. Thus, the nitro derivatives of DPA are expected to produce anions. Positive ions can also be seen from atmospheric pressure chemical ionization with nitrotoluenes [19], indicating that the nitro derivatives of DPA could also produce positive ions. Zeichner and Eldar [20] investigated the use of IMS for gunpowder residues analysis but only NG and DNT were looked for, not the stabilizers. The goal of this study was to investigate the usefulness of IMS for gunpowder stabilizers detection, which, to our knowledge, was never reported so far. This paper presents the study of positive and negative ions produced by EC, DPA and its major nitroso and nitro derivatives. In the negative ion mode, the influence of the addition of a chlorinated reactant was also investigated. Besides, some practical applications of this technique to real-world samples were also tested.

2. Experimental 2.1. Ion mobility spectrometer The ion mobility spectrometer used is a GC-IONSCAN1 M400B (Smiths Detection, USA), operated in the IMS mode. Samples can be analyzed either in the positive ion mode or in the negative ion mode. Plasmagrams were recorded using the IONSCAN System Management software, operated on a personal computer. A schematic representation of the ion mobility spectrometer is adapted from ref. [21] in Fig. 2. The sample is deposited on a sample filter and placed on the desorber heater. The sample molecules are then vapourized and carried through the heated transfer line to the ionization chamber in a flow of dry air carrier gas, possibly containing an additional reactant. The vapours are ionized by high energy electrons produced by a 63Ni beta emitter. Both positive and negative ions are formed. Depending on the mode being employed, positive or negative ions are gated (every 20 or 22 ms, with a pulse width of 0.2 ms) into the drift tube, through which they move to a collector electrode under the influence of an electric field and against a counterflow of dry air drift gas containing a calibrant. The operating conditions used are detailed in Table 1. 2.2. Chemicals Explosives, stabilizers and nitro derivatives of diphenylamine will be referred to with an abbreviation rather than by name. In Table 2, the abbreviations most often used are

94

C. West et al. / Forensic Science International 166 (2007) 91–101

Table 1 IMS operating conditions

Table 2 Abbreviations for the organic compounds used in this study

Ion mode

Parameter

Setting

Abbreviation

Compound name

Positive

Drift tube temperature (8C) Inlet temperature (8C) Desorber temperature (8C) Calibrant temperature (8C) Drift flow (mL/min) Sample flow (mL/min) Analysis time (s) Pulse duration (ms) Scan cycle time (ms) Scan number per segment Segment number per analysis

237 285 285 67 300 200 8 0.2 20 20 20

NG 2,4-DNT 2,6-DNT TNT EC

Drift tube temperature (8C) Inlet temperature (8C) Desorber temperature (8C) Calibrant temperature (8C) Drift flow (mL/min) Sample flow (mL/min) Analysis time (s) Pulse duration (ms) Scan cycle time (ms) Scan number per segment Segment number per analysis

105 240 225 63 351 300 6.6 0.2 22 20 15

Nitroglycerine 2,4-Dinitrotoluene 2,6-Dinitrotoluene 2,4,6-Trinitrotoluene Ethyl centralite (N,N’-diethyl-N,N’-diphenylurea) Diphenylamine N-Nitrosodiphenylamine 2-Nitrodiphenylamine 4-Nitrodiphenylamine 4-Nitrosodiphenylamine 2,4-Dinitrodiphenylamine 2,40 -Dinitrodiphenylamine N-Nitroso-2-nitrodiphenylamine

Negative

In the positive mode, typical operating conditions for drug detection are used. In the negative mode, typical operating conditions for explosive detection are used.

identified by IUPAC name and the structures of the molecules are represented in Figs. 1 and 3. Solvents used were HPLC grade ethanol, acetone and methylene chloride. One microgram per litre standard solutions of all compounds were prepared in ethanol. Solutions of both the individual compounds and a mixture of all compounds were prepared. All

DPA N-NODPA 2-NDPA 4-NDPA 4-NODPA 24-NDPA 240 -NDPA N-NO-2-NDPA

solutions were kept in the dark, in a refrigerator, to prevent decomposition of the nitro and nitroso derivatives [6]. The highly nitrated DPA derivatives with two or more nitro groups appear late in the decomposition process of NC and are generally of secondary interest. However, to test the applicability of the method to identify highly nitrated derivatives, we also included some dinitrated derivatives in this study. A few other compounds (NG, 2,4-DNT, 2,6-DNT and TNT, represented in Fig. 3), known to occur in combinations in regularly encountered gunpowders, were selected and analyzed as well, so as to ensure that they would not interfere with the stabilizers analysis. 2.3. Gunpowder sample preparation The method was tested with selected smokeless gunpowders, to investigate the possibility of detecting EC, DPA and its

Fig. 3. Structures of the other organic compounds analyzed. (1) TNT, (2) 2,4-DNT, (3) 2,6-DNT, (4) NG and (5) EC.

C. West et al. / Forensic Science International 166 (2007) 91–101

nitrated derivatives in real samples, where other substances might interfere. Five different balls gunpowders were analyzed. The compositions of the different gunpowders were unknown to us. The gunpowder is extracted from the cartridge, weighted and dissolved in 10 mL acetone in order to break the nitrocellulose pellets. Acetone is then evaporated to dryness and 5 mL methylene chloride are added. The residue is filtered to separate NC from the other organic components. The residue is again washed twice with 5 mL methylene chloride. Methylene chloride is a very good solvent for stabilizers and allows to separate them from NC and from inorganic salts also present in gunpowder compositions. However, chloride ions participate in the ionization chemistry in negative mode. As we wish to investigate the negative ionization chemistry with and without the chloride ion, methylene chloride must be eliminated. Thus, methylene chloride is evaporated to dryness and the residue is dissolved again in 5 mL ethanol. 2.4. Analysis procedure The sample solution was deposited in varied amounts onto a sample filter using a microlitre syringe. Solvent can be considered as a contaminant in IMS as competitive ion/molecule reactions occur that mask the response of the analyte [22]. Besides, when the solvent is present in the reaction chamber, aggregates of analyte and solvent can be formed and lead to different types of ions. Thus, throughout the experiments, a drying time of 30 s was applied, to let the solvent evaporate to dryness. All solutes were analyzed separately to determine the reduced mobility constants of their product ions formed in each operated condition, and then the mixture of all solutes was injected, to check the ability of the IMS to identify all species in the presence of possible interferents. The difficulties of quantification with the IMS are known, therefore, no quantification was attempted in this study. However, the solutes were analyzed in varied amounts to investigate the possible variations in the IMS response when different quantities of analyte are present in the spectrometer. 2.5. Reduced mobility constants and mass assignments The ion mobility constant, K (cm2 V1 s1), much as retention time in chromatography, can be used to identify the analyte from the ion peaks observed. Mobility constants are determined using the relationship: K¼

d tE

(1)

where d is the distance an ion will drift in the measured time t under the electric field E. As the ion moves through the drift gas at atmospheric pressure, under influence of the electric field, it encounters an electrostatic resistance from the drift gas molecules as well as geometric (linked to its size, shape and polarizability) and

95

diffusive forces. The following equation can be used to model these phenomena [23]:  1=2   1=2   3q 2p mþM 1 (2) K¼ 16N kT eff mM V where q is the ion charge, N the gas number density, k Boltzmann constant, Teff the effective temperature of the ion, m the ion mass, M the neutral gas molecule mass and V is the collision cross-section of the ion. The latter is determined by ionic size, shape, symmetry and charge distribution [24]. For the purpose of standardization, as absolute mobility varies as a function of drift gas density, ion mobility reduced to standard temperature and pressure, K0, is preferably reported:    273 P K0 ¼ K (3) T 101; 325 where T is the temperature (in K) and P is the pressure (in kPa). In practice, a reference ion (the ‘‘calibrant’’) is used to determine K0. The calibrant mobility is used to correct daily K0 values. Using reduced mobilities from the litterature [21], a graph of 1/K0 versus ion mass was constructed, from which approximate molecular weights of the ions observed could be determined. Then proposals were made for the identities of the ions. Mass assignments based upon mobility are in general only accurate to 20% [25]. The exact nature of the species formed in the IMS can only be decisively determined using IMS–MS combination. Thus, the identities suggested for the varied ions observed should be regarded as speculative. 3. Results and discussion 3.1. Positive ion mode In the positive mode of the IONSCAN, the drift gas contains trace amounts of nicotinamide (NTA) used as both calibrant and reactant. The analyte molecule gets ionized, according to the following proton transfer reaction: ½NTAHþ þ M ! NTA þ MHþ

(4)

This reaction only proceeds if the proton affinity of the sample molecule is greater than that of NTA [22]. Ion lifetimes must also be considered as variables for optimum response or resolution [16]. Instances where response is not observed for an analyte can be attributed to ion instability or low proton affinity. The reduced mobility constants (K0) of the dominant peaks for each compound were calculated and are listed in Table 3. EC and DPA were both seen to produce one well-resolved ion peak. The ion produced by EC is apparently thermally and chemically very stable as the amplitude of the peak is largely higher than that of DPA analyzed in identical quantities. Fig. 4a and b show their product ion spectra, also called plasmagrams. The identity of the response ions has not been determined by coupling the ion mobility spectrometer with a mass spectrometer, but the reduced mobility is consistent with that of the

96

C. West et al. / Forensic Science International 166 (2007) 91–101

Table 3 Reduced mobilities (K0) and proposed identities of the observed ions obtained when stabilizers are analyzed in positive ion mode Compound

MW

K0 (cm2 V1 s1)

Estimated MW

Proposed identity +

Proposed MW

EC DPA

268 169

1.24 1.61

267 152

ECH DPAH+

269 170

N-NODPA

198

1.08 1.03

345 368

(M–NO)(M–NO)H+ (M–NO)(M–NO)NO+

337 366

2-NDPA 4-NDPA 4-NODPA

214 214 198

1.46 – 1.47

190

MH+

215

188

MH+

199 +

N-NO-2-NDPA

243

1.47 1.52

189 175

(M–NO)H (M–NO2)H+

214 198

24-NDPA 240 -NDPA

259 259

– 1.47

189

(M–NO2)H+

214

MW is the mass of the neutral molecule Estimated MW is the molecular weight estimated from the reduced mobility. Proposed MW is the molecular weight of the proposed identity.

Fig. 4. Positive ion spectra of: (a) EC, (b) DPA and (c) N-NODPA.

C. West et al. / Forensic Science International 166 (2007) 91–101 Table 4 Reduced mobilities (K0) of the observed ions obtained when stabilizers are analyzed in negative ion mode, with hexachloroethane Compound

MW

K0 (cm2 V1 s1)

Estimated MW

EC DPA N-NODPA 2-NDPA 4-NDPA 4-NODPA

268 169 198 214 214 198

– – – – 1.24 1.27

315 303

N-NO-2-NDPA

243

1.24 1.17

317 353

24-NDPA

259

1.25

311

240 -NDPA

259

1.18 1.11

351 390

MW is the mass of the neutral molecule Estimated MW is the molecular weight estimated from the reduced mobility.

nondissociated product ion, ionized through a proton attachment process. Indeed, judging by the calculated mass of the ions produced by EC and DPA, it is reasonable to assume that these are the molecular ions. Extensive quantitative investigations of response factors have not yet been attempted, although the approximate limit of detection for EC was determined to be in the range of 0.5–1 ng; for DPA, it was approximately 2 ng. The low mobility of the two ions produced by N-NODPA (Fig. 4c) suggests the formation of dimers. Actually, N-NODPA is known to denitrosate during thermospray mass spectrometry at 250 8C [4]. A DPA radical is formed as a result. Then two DPA radicals can combine to form tetraphenylhydrazine (MW 337). This dimer (presenting a reduced mobility K0 = 1.075 cm2 V1 s1), can only appear when large amounts of N-NODPA are introduced in the IMS. The second ion (with a reduced mobility K0 = 1.033 cm2 V1 s1) observed may be a dimer-adduct of some sort, possibly a tetraphenylhydrazine-nitroso adduct. The other nitro and nitroso derivatives of DPA apparently do not face such a reaction, as no other heavy ion is observed. Therefore, the presence of the two dimer peaks is diagnostic of N-NODPA. The identity of the cations produced by the nitro and nitroso derivatives of DPA is not clear. Two types of ions seem to be formed: one with a mobility constant close to 1.47 cm2 V1 s1, and the other with a mobility constant of 1.52 cm2 V1 s1.

97

According to the mass calculations, degraded product ions can be suggested for some of them. The gas phase basicity is related to the proton affinity and to the three-dimensional structure of the analyte [26]. Therefore, the isomers of a compound can form different types of ions because of their geometrical structure, or their internal charge distribution, or both. Furthermore, these characteristics can also induce different mobilities for isomeric species. As a result, 4-NDPA and 2,4-NDPA do not produce any visible cation, although their isomeric forms 2-NDPA and 2,40 -NDPA do. 3.2. Negative mode 3.2.1. With chlorinated reactant An important variation of IMS ionization, that can be used to enhance the sensitivity and selectivity of the technique for particular classes of compounds, or to simplify the response for certain analytes, involves the addition of reactants to the drift gas. In the negative ion mode, the internal calibrant in IONSCAN is 4nitrobenzonitrile and the reactant is hexachloroethane. The hexachloroethane reactant yields alternate reactant ions such as Cl, NO3 or NO2. Consequently, ions formed with explosives in IMS occur through APCI reactions using Cl, NO3 or NO2 [19]. NO3 or NO2 are generated from the nitrated solute itself and are consumed in the ion source through association with other solute molecules. These processes can be described by Eqs. (5)–(7): e þ M ! NOx  þ ðMNOx Þ

(5)

NOx  þ M ! MNOx 

(6)

Cl þ M ! MCl

(7)

EC and DPA do not produce any visible negative ion. They might not have any proton acidic enough to be subtracted. In most of DPA derivatives, the presence of electro-attracting nitro and nitroso groups on the aromatic ring apparently enhances the acidic character and allows the formation of stable anions (see Table 4). An example is given with the spectrum produced by 2,4-NDPA (Fig. 5). Again the isomeric forms show different behaviours: 2-NDPA does not produce any negative ion, while its isomeric

Fig. 5. Negative ion spectra of 2,4-NDPA, obtained with the chlorinated reactant.

98

C. West et al. / Forensic Science International 166 (2007) 91–101

form 4-NDPA does; N-NODPA does not produce any negative ion, while its isomeric form 4-NODPA does; the di-nitro isomers 2,4-NDPA and 2,40 -NDPA produce peaks with different mobilities. Often, the gas phase chemistry of nitrotoluenes can be described simply as the formation of M or (M–H), depending upon the availability of a proton abstracting reagent ion (such as Cl) and the acidity of protons. Thus, nitrated derivatives of DPA were expected to produce molecular negative ions. However, the calculated molecular weights of the ionic species formed are higher than what would be expected on the basis of molecular ion structure. Therefore, the formation of adducts with chlorine – originating from the hexachloroethane reactant – and nitro groups is probable.

Table 5 Reduced mobilities (K0) of the observed ions obtained when stabilizers are analyzed in negative ion mode, without hexachloroethane Compound

MW

K0 (cm2 V1 s1)

Estimated MW

EC DPA N-NODPA 2-NDPA

268 169 198 214

– – – 1.34

271

4-NDPA

214

1.34 1.24 1.17 1.13 1.09

270 317 352 377 401

47 35 25 24

1.37 1.27 1.20 1.15 1.12

257 302 338 364 386

45 36 26 22

1.34 1.25 1.17 1.13 1.09

270 315 352 377 401

45 37 25 24

1.37 1.27 1.24 1.20 1.15 1.11

257 303 315 338 364 388

46 12 23 26 24

1.34 1.24 1.18 1.11

270 317 349 389

47 32 40

4-NODPA

198

3.3. Without chlorinated reactant From our own experience of explosive analysis in IMS, the chlorine reactant ion chemistry is not always able to resolve peak overlap of analytes. This point was also observed by Daum et al. [27], who showed that the resolution of analytes can sometimes be achieved using only purified air for the formation of reactant ions. In particular, we observed a lower detection limit for dinitrotoluenes, possibly due to a higher stability of the ions formed in the absence of the chlorinated reactant. Furthermore, the different isomers of DNT could be discriminated as their mobilities were slightly different, while they are all identical when hexachloroethane is used. Thus, for particular compounds, both selectivity and sensitivity could be enhanced by this mean. For this reason, we investigated the ionization occuring when suppressing the hexachloroethane reactant in negative mode. The results are presented in Table 5. As was already observed when hexachloroethane was used as a reactant, EC, DPA and N-NODPA do not produce any visible negative ion. Again, the isomeric species show different behaviours: 2-NDPA produces a unique ion (see Fig. 6a) while 4-NDPA produces five ions, one of them being very close to the 2-NDPA ion; 4-NODPA produces five ions (see Fig. 6b) while N-NODPA produces none; 2,4-NDPA produces six ions while 2,40 -NDPA produces three ions. The peaks produced by N-NO-2-NDPA, 4-NODPA, 4NDPA, 2,4-NDPA and 2,40 -NDPA seem to be produced by the formation of similar adducts, as the difference in the calculated mass between consecutive ions (DMW in Table 5) are nearly identical. Therefore, the clustered groups may be of the same nature, but the identity of the respective ions remains in question. It has to be emphasized that, in this particular case, coupling the ion mobility spectrometer to a mass spectrometer in an attempt to identify the product ions might be of no help as the adduct ions might not survive the transition between the high ambient pressure and the very low pressure necessary for the MS analysis: weakly bound cluster ions might be collisionally decomposed in the interface region. The IONSCAN uses purified ambient air, which may contain a variety of chemicals, for both drift gas and carrier

N-NO-2-NDPA

24-NDPA

240 -NDPA

243

259

259

DMW

MW is the mass of the neutral molecule estimated MW is the molecular weight estimated from the reduced mobility. DMW is the difference in the estimated molecular weight of two consecutive ion peaks.

gas. Indeed, room air commonly produces three to five reactant ions [28]. Many ion/molecule reactions will occur within the source residence time. The reactant ions in negative polarity are thermalized electrons in nitrogen and are hydrated O2 and CO2 [19]. But, if the solvent is not well evaporated before the analysis, clusters formed with solvent molecules (ethanol, in this case) must also be considered. When high-temperature operation with counter current drift-gas flow is employed, humidity effects from air samples introduced into the IMS are normally not observed [29]. Thus, in the present operating conditions, water molecules aggregates should not be observed. Increasing the temperature in the drift tube might eliminate some of the clusters formed, thus favouring a simpler ion mobility spectrum, but the temperature conditions are supposedly optimized for explosive detection, therefore no attempt was made at varying the temperature. Whatever the reason for these multiple peaks, and although some peaks for all these species are not clearly separated due to similar drift times, the IMS could still discriminate all these analogs when the species were analyzed in a mixture.

C. West et al. / Forensic Science International 166 (2007) 91–101

99

Fig. 6. Negative ion spectra of: (a) 2-NDPA and (b) 4-NODPA, obtained without the chlorinated reactant.

3.4. Establishing a detection algorithm The responses observed for all the substances evaluated in this paper in all modes of operation tested are reported in Table 6. It can be observed that any substance tested can be analyzed with IMS, be it in the positive or the negative ion mode, with or without the chlorinated reactant. It is also of value that 2-NDPA and 4-NDPA, being some of the primary nitro derivatives formed in the gunpowder composition, are detected in two different modes of operation, resulting in a better selectivity after a positive detection in both modes. In any operating mode, when chemical identification of the ions observed was suggested, it cannot be more than speculative. However, the measured values of mobility are characteristic of the sample material and can be used for Table 6 Summary of the response observed when stabilizers are analyzed in positive and negative ion modes, with and without chlorinated reactant Compound

EC DPA N-NODPA 2-NDPA 4-NDPA 4-NODPA N-NO-2-NDPA 24-NDPA 240 -NDPA

Positive mode

Negative mode With Cl

Without Cl

+ + + + +

+ + + + + +

+ + + + + + +

identification in the same manner as retention times are used in chromatography. Nevertheless, a confirmation of the identity of the substance, provided by an orthogonal method, would be required for positive identification. For instance, coupling the IMS after a chromatograph would provide additional information (the chromatographic retention time) about the substance detected by the IMS. By IMS, the production of several peaks for one substance can be helpful. In the presence of an interfering species, unless all peaks were affected, the substance could still be identified. Moreover, multiple peaks reduce the false alarm rate. However, multiple peaks reduce the sensitivity since the species are distributed among more than one ion and quantification is more difficult. In our case, precise qualitative information is preferable to quantitative information. For instance, when analysing the stabilizers mixture in negative mode with the chlorinated reactant, the ion peak produced by 4-NDPA (K0 = 1.24 cm2 V1 s1) could be mistaken for an ion peak produced by N-NO-2-NDPA, having a close mobility (K0 = 1.24 cm2 V1 s1). However, N-NO-2-NDPA also produces a second ion peak with clearly different mobility (K0 = 1.17 cm2 V1 s1). Thus, these two species can be differentiated. Furthermore, the relative intensities of the multiple ion peaks are influenced by concentration effects and by progress through the IONSCAN desorption cycle. Detection algorithms were then developed to take account of the multiple-peak nature of the nitro derivatives of DPA and to accommodate their varying relative intensities and the different patterns occurring when the concentration of the

100

C. West et al. / Forensic Science International 166 (2007) 91–101

Table 7 Summary of the substances identified in gunpowder sample #4 analyzed in positive and negative ions modes, with and without the chlorinated reactant Compound

EC DPA N-NODPA 2-NDPA 4-NDPA N-NO-2-NDPA NG TNT DNT 2,4-DNT 2,6-DNT

Positive mode

Negative mode With Cl

Without Cl

+ + + + +

+ + +

+ + +

+ +

substance is varied. Thus, the IMS can be used to create a stabilizer profile of the smokeless gunpowder. 3.5. Gunpowder samples Each of the selected powders yielded suitable plasmagrams. No interference was observed. The original premise that IMS could be used to analyze smokeless powder was then confirmed. 4-NODPA was never detected, nor were the di-nitro derivatives 2,4-NDPA and 2,40 -NDPA. This may be because 4-NODPA is a relatively reactive and unstable compound and the gunpowders may be in the early stages of decomposition, where di-nitro derivatives only begin to form. This also confirms that the IMS method and the detection algorithms developed are able to differentiate between the different nitro and nitroso derivatives, although their product ions may sometimes be very close on the ion mobility spectrum. Thus, it is possible to selectively detect a single DPA derivative in a complex mixture. An example is given in Table 7, where the results obtained with sample #4 are presented. It is clear that the use of the positive and negative modes in conjunction is most informative on the composition of the gunpowder. In this case, the analysis performed in negative ion mode without the use of hexachloroethane does not bring any fundamental information, apart from the more precise identification of DNT isomers. In other cases, this point proved to be useful as the detection limit for DNT is lower and DNT could then be detected in the negative ion mode without chlorine reactant, while it was not detected in the negative ion mode when the chlorinated reactant was present. 4. Conclusion The data obtained in the present study show that the method employed could play a valuable role in the definitive analysis of gunpowders in casework samples. Analyses are fast, inexpensive and generate a minimum amount of waste. Moreover, the joint detection of several components of smokeless gunpowder

reduces the possibility of interferences and the identification of gunpowder is therefore more definitive. In comparison to traditionally used TLC method, the technique is two to three orders of magnitude more sensitive. Compared to HPLC with diode-array detection, it is still one order of magnitude more sensitive. Dual-mode detection would allow an operator to screen for EC, DPA and N-NODPA in the positive mode while screening for NG, TNT, DNT and the nitro derivatives of DPA in the negative mode. Other components possibly found in gunpowder compositions could be analyzed by this technique. For instance, the method is surely also suitable for the analysis of methylcentralite, since its structure is so similar to that of EC. Besides, phthalate compounds were reported to produce positive ions [18]. Moreover, Kuja et al. [30] reported the detection of NC in negative ion mode. Therefore, if positive and negative ions were monitored at the same time, as is the case with Ionscan 500DT, a complete analysis of smokeless gunpowder organic compounds could be achieved in one single IMS analysis, in less than 10 s. Nevertheless, for any casework sample, a confirmation provided by an orthogonal method is compulsory. In this respect, the very small amounts of sample required by the IMS analysis are an advantage, as it possibly leaves sample for complementary analyses. References [1] Y. Tong, Z. Wu, C. Yang, J. Yu, X. Zhang, S. Yang, X. Deng, Y. Xu, Y. Wen, Determination of DPA in smokeless gunpowder using a tandem MS method, Analyst 126 (2001) 480–484. [2] J. Yinon, A. Acevedo, T. Chamberlain, S. Brunk, Differentiation between nitroglycerin explosive and nitroglycerin medication using an IMS detector, in: D. Garbutt, P. Pilon, P. Lightfoot (Eds.), Proceedings of the 8th International Symposium on Analysis and Detection of Explosives, 2004, pp. 306–313. [3] O. Drzyzga, Diphenylamine and derivatives in the environment, Chemosphere 53 (2003) 809–818. [4] J.C. Via, L.T. Taylor, Chromatographic analysis of nonpolymeric single base propellant components, J. Chromatogr. Sci. 30 (1992) 106–110. [5] A. Bergens, R. Danielsson, Decomposition of diphenylamine in nitrocellulose based propellants I. Optimization of a numerical model to concentration-time data for diphenylamine and its primary degradation products determined by liquid chromatography with dual amperometric detection, Talanta 42 (1995) 171–183. [6] J.M. Bellerby, M.H. Sammour, Stabilizer reactions in double base rocket propellants, propellants, explosives, Pyrotechnics 16 (1991) 235–239. [7] A. Zeichner, B. Eldar, B. Glattstein, A. Koffman, T. Tamiri, D. Muller, Vacuum collection of gunpowder residues from clothing worn by shooting suspects, and their analysis by GC/TEA, IMS and GC/MS, J. Forensic Sci. 48 (5) (2003) 961–972. [8] O. Cascio, M. Trettene, F. Botolotti, G. Milana, F. Tagliaro, Analysis of organic components of smokeless gunpowders: HPLC vs. Micellar electrokinetic capillary chromatography, Electrophoresis 25 (2004) 1543–1547. [9] A. Bergens, Decomposition of diphenylamine in nitrocellulose based propellants II. Application of a numerical model to concentration–time data determined by liquid chromatography and dual-wavelength detection, Talanta 42 (1995) 185–196. [10] J.B.F. Lloyd, Liquid chromatography of firearms propellant traces, J. Energetic Mater. 4 (1986) 239–271.

C. West et al. / Forensic Science International 166 (2007) 91–101 [11] H. Meng, B. Caddy, Detection of N,N0 -diphenyl-N,N0 -diethylurea (ethylcentralite) in gunshot residues using high-performance liquid chromatography with fluorescence detection, Analyst 120 (1995) 1759–1762. [12] Z. Wu, Y. Tong, J. Yu, X. Zhang, C. Pan, X. Deng, Y. Xu, Y. Wen, Detection of N,N0 -diphenyl-N,N0 -dimethylurea (methyl centralite) in gunshot residues using MS–MS method, Analyst 124 (1999) 1563–1567. [13] M. Ashraf-Khorassani, L.T. Taylor, Qualitative supercritical fluid chromatography/Fourier transform infrared spectroscopy study of methylene chloride ans supercritical carbon dioxide extracts of double-base propellant, Anal. Chem. 61 (1989) 145–148. [14] Y. Tong, Z. Wu, C. Yang, J. Yu, X. Zhang, S. Yang, X. Deng, Y. Xu, Y. Wen, Determination of diphenylamine in smokeless gunpowder using a tandem MS method, Analyst 126 (2001) 480–484. [15] G.A. Eiceman, Z. Karpas, Ion Mobility Spectrometry, CRC Press, Boca Raton, FL, 1994. [16] G.A. Eiceman, Ion mobility spectrometry as a fast monitor of chemical composition, Trends Anal. Chem. 21 (2002) 259–275. [17] Z. Karpas, Ion mobility spectrometry of aliphatic and aromatic amines, Anal. Chem. 61 (1989) 684–689. [18] G. Simpson, M. Klasmeier, H. Hill, D. Atkinson, G. Radolovich, V. LopezAvila, T.L. Jones, Evaluation of gas chromatography coupled with ion mobility spectrometry for monitoring vinyl chloride and other chlorinated and aromatic compounds in air samples, J. High Resolut. Chromatogr. 19 (1996) 301–312. [19] R.D. Ewing, D.A. Atkinson, G.A. Eiceman, G.J. Ewing, A critical review of ion mobility spectrometry for the detection of explosives and explosive related compounds, Talanta 54 (2001) 515–529. [20] A. Zeichner, B. Eldar, A new method for extraction and analysis of gunpowder residues on double-side adhesive coated stubs, J. Forensic Sci. 49 (2004).

101

[21] Anonymous, GC-IONSCAN1 Model 400B Operator’s Manual, Smiths Detection, USA. [22] T. Keller, A. Miki, P. Regenscheit, R. Dirnhofer, A. Schneider, H. Tsuchihashi, Detection of designer drugs in human hair by ion mobility spectrometry, Forensic Sci. Int. 94 (1998) 55–63. [23] H.E. Revercomb, E.A. Mason, Theory of plasma chromatography/gaseous electrophoresis—a review, Anal. Chem. 47 (1975) 970–983. [24] N. Agbonkonkon, H.D. Tolley, M.C. Asplund, E.D. Lee, M.L. Lee, Prediction of gas-phase reduced ion mobility constants (Ko), Anal. Chem. 76 (2004) 5223–5229. [25] G.W. Griffin, I. Dzidic, D.I. Carroll, R.N. Stillwell, E.C. Horning, Ion mass assignments based on mobility measurements, Anal. Chem. 45 (1973) 1204–1209. [26] Y. Guo, M.Q. Lu, Y.T. Long, Ion mobility spectra of selected amines and their application in field testing with the use of a portable IMS device, Field Anal. Chem. Technol. 1 (4) (1997) 195–211. [27] K.A. Daum, D.A. Atkinson, R.G. Ewing, W.B. Knighton, E.P. Grimsrud, Resolving interferences in negative mode ion mobility spectrometry using selective reactant ion chemistry, Talanta 54 (2001) 299–306. [28] P. Rodacy, P. Leslie, S. Klassen, R. Silva, Ion mobility spectroscopic techniques for the detection and identification of explosives, in: C.R. Midkiff (Ed.), Proceedings of the 5th International Symposium on Analysis and Detection of Explosives, 1997. [29] H.H. Hill, G. Simpson, Capabilities and limitations of ion mobility spectrometry for field screening applications, Field Anal. Chem. Technol. 1 (3) (1997) 119–134. [30] F. Kuja, A. Grigoriev, R. Debono, S. Nacson, Ion mobility spectrometry in the detection of improvised explosives, in: A. Cumming (Ed.), Proceedings of the 7th International Symposium on Analysis and Detection of Explosives, 2001, pp. 179–184.

Forensic Science International 166 (2007) 102–109 www.elsevier.com/locate/forsciint

A study of the use of Ephedra in the manufacture of methamphetamine W.D. Barker a,*, U. Antia b a

Institute of Environmental Science and Research Ltd (ESR), Mt Albert Research Centre, Hampstead Road, Private Bag 92021, Auckland, New Zealand b University of Auckland, Department of Chemistry, Auckland, New Zealand Received 20 February 2006; received in revised form 2 April 2006; accepted 9 April 2006 Available online 16 May 2006

Abstract The Ephedra plant has been identified as an excellent source of ephedrine and pseudoephedrine, both of which can be chemically reduced to form the widely abused illicit drug methamphetamine. Ephedra contains several additional alkaloids that undergo analogous reductions to form amphetamine and N,N-dimethylamphetamine (also drugs of abuse). The main alkaloids obtained from the Ephedra plant have been reduced using four common methods used by the clandestine operator. The intermediates and byproducts of these reductions have been identified and/or tentatively assigned and the mechanism of formation discussed. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Ephedra; Methamphetamine; Amphetamine; Dimethylamphetamine; Reduction; Intermediate; Byproduct

A frequently encountered method for the clandestine manufacture of methamphetamine involves the conversion of pseudoephedrine and/or ephedrine to methamphetamine by reduction [1]. Pseudoephedrine and ephedrine are commonly obtained from pharmaceutical preparations, which are often available from drug stores or pharmacies (depending on local legislation). An alternative source of ephedrine and pseudoephedrine is the naturally occurring plant Ephedra. Ephedra is a primitive stalky plant that contains numerous alkaloids including ephedrine and pseudoephedrine. The ground up plant material (also referred to as Ma Huang) is frequently seen in tablet, capsule or powdered form. Ma Huang is a Chinese herbal remedy used to relieve respiratory related ailments such as bronchitis and asthma [2]. There have been more than 30 different species of Ephedra found, mainly in subtropical and temperate regions of Europe, Asia and America. However, only a few of these species contain ephedrine related alkaloids at any significant level [3]. The main alkaloids present in Ephedra are the physiologically active diastereomeric pairs ()-ephedrine and (+)-pseudoephedrine; ()-methylephedrine and (+)-methylpseudoephedrine;

* Corresponding author. Tel.: +64 98153949. E-mail address: [email protected] (W.D. Barker). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.005

()-norephedrine and (+)-norpseudoephedrine (Fig. 1). The enantiomers of the six compounds in Fig. 1 are not physiologically active and have not been observed in nature. Previous quantitative studies show that alkaloid levels vary widely inter and intra species and in general ephedrine and pseudoephedrine are more abundant than norephedrine and norpseudoephedrine which are, in turn, more abundant than methylephedrine and methylpseudoephedrine [2]. Locally sourced Ephedra americana and Ephedra campylpoda were found to contain approximately 1% total alkaloid (dry weight).1 Fig. 1 clearly illustrates that the alkaloids differ only by the alkylation of the amine and this is carried through in the main reduction product of each pair of diastereoisomers. As expected the tertiary amine pair ()-methylephedrine and (+)-methylpseudoephedrine yield (+)-N,N-dimethylamphetamine, the secondary amine pair ()-ephedrine and (+)-pseudoephedrine yield (+)-methamphetamine and the primary amine pair ()norephedrine and (+)-norpseudoephedrine yield (+)-amphetamine [4]. From a forensic chemist’s perspective it is the presence of each of these compounds in an illicit sample that can provide information on the original source of the sample. For example,

1 The extraction of these alkaloids can be variable depending on technique and solvents used (U. Antia, unpublished results).

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

103

Fig. 1. The six physiologically active ephedrine related alkaloids found in Ephedra and their three reduction products.

if a methamphetamine product contains a small amount of amphetamine and N,N-dimethylamphetamine then Ephedra is the likely precursor in the manufacturing process [5]. A further tool available to the forensic chemist is the identification of other byproducts, reaction intermediates and impurities in samples associated with the clandestine methamphetamine manufacturing process. The detection of these additional compounds within a sample often provides information on the synthetic route used to produce a particular product. For example, cis- and trans-1,2-dimethyl-3-phenylaziridine, 1-phenyl-2-propanone, 1,3-dimethyl-2-phenylnaphthalene and 1-benzyl-3-methylnaphthalene are route specific intermediates and byproducts of the manufacture of methamphetamine using any of the many variations of the hydriodic acid reduction of pseudoephedrine/ephedrine (Fig. 2) [6–8]. This article expands upon previous work by investigating several frequently observed methods of clandestine methamphetamine manufacture using each of the six main Ephedra alkaloids as a precursor. Intermediates and byproducts have been identified and evaluated providing the forensic chemist with additional information for the investigation of clandestinely produced drugs. 1. Materials and instrumentation Anhydrous ammonia gas was purchased from BOC Gases. Pseudoephedrine, norephedrine and norpseudoephedrine were purchased from Acros Organics. Lithium metal was obtained

from lithium AA batteries (Energizer). Methylephedrine and methylpseudoephedrine were manufactured in the laboratory from ephedrine and pseudoephedrine using previously published syntheses and characterized prior to use [9]. Ephedrine hydrochloride was a seized sample and was characterized prior to use using authenticated standards. Thionyl chloride was purchased from Scharlau Chemie. Palladium chloride was also seized but originated from Kee Shing Industrial Products. Barium sulfate was purchased from Panreac Quimica SA. All other chemicals and solvents were purchased from BDH Laboratory Supplies. Gas chromatography–mass spectrometry (GCMS) analysis was conducted using an Agilent 6890N Network Gas Chromatograph with a 5973N inert Mass Selective Detector. A 25 m BPX5 220 mm i.d. column with a 0.25 mm film thickness was used with helium carrier gas. After 2 min at 70 8C, the temperature was ramped to 300 8C at 30 8C/min. The samples were prepared by extraction into chloroform. Nuclear magnetic resonance (NMR) was conducted using a Bruker Biospin AVANCE DRX 400 spectrometer at 400.17 MHz for proton NMR and 100.61 MHz for carbon NMR. The samples were dissolved in deuterated chloroform for analysis. The following reductions were carried out using variations of previously reported methods: red phosphorus and iodine reduction [6,7,8], hypophosphorous acid and iodine reduction [10], dissolving metal reduction [11–13] and metal catalysed reduction [14,15]. Specific details of the reduction reactions and analytical data are available from the author.

Fig. 2. Route specific intermediates and byproducts of the hydriodic acid reduction of pseudoephedrine/ephedrine [6–8].

104

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

2. Results and discussion The four frequently encountered reduction reactions investigated were two variations of the hydriodic acid reduction, a dissolving metal reduction and a metal catalysed hydrogenation (via a chloro-intermediate). All four reductions were carried out on each of the six main Ephedra alkaloids and the reaction progress monitored by GCMS. Acidic and basic extracts of the reaction mixtures were taken at time intervals to evaluate the maximum number of byproducts and intermediates.

2.1. Hydriodic acid reduction Hydriodic acid can be used to directly reduce benzyl alcohols under reflux conditions, however, during the clandestine manufacture of methamphetamine it is more common to proceed via the in situ generation of the reactive ‘‘HI’’ species. The reduction is believed to progress via an iodo intermediate of the alkaloid by displacement of the hydroxyl group with the iodo anion [1]. While the iodoephedrine (or equivalent) is never isolated within the reaction mixture, previous mechanistic studies add significant weight to this hypothesis [6–8]. Two frequently encountered methods for ‘‘HI’’ production involve mixing iodine with red phosphorus in the presence of water or mixing hypophosphorous acid with iodine. Both of these methods were used during this investigation. The only difference in results was the time scale of intermediate, byproduct and product formation. The hypophosphorous acid/iodine reduction was significantly faster than the red phosphorus/iodine reduction. As discussed previously, the reduction of ephedrine by hydriodic acid generates a number of intermediates and byproducts (Fig. 2), which can be observed forming during a reaction. Previous studies show that as the reduction progresses, ephedrine is consumed while the aziridines (cis- and trans-1,2dimethyl-3-phenylaziridine) evolve. Over time a new species 1phenyl-2-propanone (P-2-P) forms, as does methamphetamine. Towards the end of the reaction time scale, the aziridines disappear and the naphthalenes (1,3-dimethyl-2-phenylnaphthalene and 1-benzyl-3-methylnaphthalene) begin to form [6–8]. It is well accepted that the aziridine intermediates2 are formed by elimination of iodide from the iodo intermediate. During the progress of the reaction, the aziridines are consumed as they are either reduced to methamphetamine or hydrolysed to form P-2-P. Over time, P-2-P dimerises via an acid catalysed intermolecular condensation to give the naphthalenes. Pseudoephedrine undergoes similar reduction resulting in the same byproducts and intermediates [6–8].

Fig. 3. Reaction progress of the red phosphorus/iodine reduction of norephedrine (base-ether extraction of an aliquot of reaction mixture at t = 30 min, analysed by GCMS).

Norephedrine reduced to amphetamine under the same conditions. The reaction progressed in a similar manner and gave rise to the expected intermediates (Fig. 3). In this instance, the aziridines observed were, as expected, cis- and trans-3-phenyl-2-methylaziridine. Small amounts of P2-P were observed in low levels in an acid extraction of the reaction mixture, however, during chromatography P-2-P coelutes with amphetamine and is therefore not readily identified. As the reaction progressed, the aziridines were again consumed while the P-2-P condensation products were formed (Fig. 4). An additional peak attributed to the formation of N(b-phenylisopropyl)benzyl-methylketimine was observed at approximately 8.5 min. The ketimine is likely to be a condensation product of P-2-P and the primary amine of amphetamine. Norpseudoephedrine reduced to amphetamine under the same conditions and gave rise to the same byproducts/ intermediates. Methylephedrine reduced to N,N-dimethylamphetamine in a similar manner with the following exception. Basic extracts of the reaction mixture over time were of limited value as they exhibited starting material and product with no additional compounds observed. Acidic extracts (direct ether extract of an aliquot of reaction mixture in water) did not exhibit the aziridine intermediates, however, they did display two new tentatively identified ‘‘intermediates’’: 1-propenylbenzene and 2-propenylbenzene (Fig. 5). The presence of P-2-P in the reaction mixture indicates that aziridines were likely to have formed at some stage, however, they were not observed during the GCMS analysis.

2

While it is not fully agreed that the aziridine species are actually formed during the reduction, they are commonly observed during chromatographic analysis. The actual intermediate responsible for the observed aziridines may be an iminium ion, an enaminium ion, an aziridinium ion or a combination of all three species [6,7,8].

Fig. 4. Reaction progress of the red phosphorus/iodine reduction of norephedrine (base-ether extraction of an aliquot of reaction mixture at t = 90 min, analysed by GCMS).

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

105

Fig. 5. Reaction progress of the red phosphorus/iodine reduction of methylephedrine (acid extraction at t = 15 min, analysed by GCMS).

The predicted aziridine intermediates would be quaternary N-alkylated cis- and trans-1,1,2-trimethyl-3-phenylaziridine. The instability of these charged intermediates during chromatographic conditions may lead to the formation of the benzylpropenes by elimination. The fact that during the progress of the reaction, the benzylpropenes disappear adds weight to this hypothesis. If the benzylpropenes were in fact formed during the reaction, as relatively unreactive byproducts (rather than eliminated intermediates), they should remain unchanged in the final reaction mixture. Methylpseudoephedrine also reduced to N,N-dimethylamphetamine under the same conditions and gave rise to the same byproducts/intermediates. 2.2. Dissolving metal reduction The dissolving metal reduction reaction was developed during the 1940s as a method for synthesising cyclohexadienes from arenes [16]. More recently, the utility of lithium/ammonia in the selective reduction of benzyl alcohols has been recognised [17]. This has subsequently led to the technique being used extensively in the clandestine manufacture of methamphetamine [11–13]. The reduction of the Ephedra alkaloids to their related amphetamines has been previously studied and is believed to occur via an electron-mediated process leading to the heterolytic cleavage of the hydroxyl group [11]. A limitation of this reaction is the over-reduction of the amphetamine product when an excess of the alkali metal is present in the reaction mixture. For example samples of methamphetamine produced in a clandestine environment often contain a byproduct, which has recently been identified as 1-(10 ,40 cyclohexadienyl)-2-methylaminopropane (CMP) [12,18].

Fig. 6. Mass spectra of CMP (byproduct formed during the manufacture of methamphetamine by reduction using the lithium ammonia method).

Fig. 7. Mass spectra of byproducts of norephedrine/norpseudoephedrine reduction (top) and methylephedrine/methylpseudoephedrine reduction (bottom) and tentatively assigned structures.

The protons required for the reduction of the hydroxyl group and the partial reduction of the aromatic ring arise because of damp or impure solvents or even water absorbed into the reaction from the atmosphere. CMP is another route specific byproduct and is indicative of a methamphetamine sample being synthesised by this method. Reproduction of the reduction of pseudoephedrine and ephedrine using a previously published method resulted in methamphetamine and a small amount of the CMP byproduct. The mass spectra of CMP, observed in both reduction reactions, was consistent with the literature (Fig. 6) [12,18]. Norephedrine and norpseudoephedrine were reduced to amphetamine using the same conditions, while methylephedrine and methylpseudoephedrine were reduced to N,Ndimethylamphetamine. In each reaction, a minor byproduct was observed. The byproducts were tentatively identified by comparison to the previously characterized reduction product CMP (Fig. 7).3 The parent ion of the amphetamine byproduct {m/z 136 (M  1)}, is 14 mass units (a methylene group) lower than that observed for CMP, thus indicating the CMP analogue 1-(10 ,40 cyclohexadienyl)-aminopropane (CAP). Conversely the parent ion in the N,N-dimethylamphetamine byproduct {m/z 164 (M  1)}, is 14 mass units higher, indicating the CMP analogue 1-(10 ,40 -cyclohexadienyl)-2,2-dimethylaminopropane (CDP). The methylephedrine and methylpseudoephedrine reacted much slower than the less hindered analogues and over the time period of the reaction there was a significant amount of unreacted starting material left. 3 A further minor byproduct previously hypothesised as a CMP ring isomer was also observed at low levels in the reduction of pseudoephedrine and ephedrine to methamphetamine. An analogous byproduct was also seen in the amphetamine and dimethylamphetamine products. However, due to the low relative concentration, the analytical data associated with these compounds is of limited value.

106

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

2.3. Metal catalysed hydrogenation (via chlorointermediate) The benzylic alcohol moiety present in the Ephedra alkaloids cannot be effectively reduced using hydrogenation techniques commonly available to the clandestine chemist. Instead, ephedrine and pseudoephedrine are generally converted to their chloro-derivatives prior to reduction by catalytic hydrogenation. Previous work in the area shows that the chloroderivatives of ephedrine and pseudoephedrine can be synthesised in relatively high yield using a one step process, involving thionyl chloride or phosphorus pentachloride [14,15,19]. There appears to be some disagreement within the literature as to the nature of the chloro-product as Allen and Kiser [14] describe ()-ephedrine leading almost exclusively to (+)-chloropseudoephedrine (99%) via an SN2 process and (+)-pseudoephedrine leading to a 60:40 mixture of (+)-chloropseudoephedrine and ()-chloroephedrine via a combination of SN2 and SNi displacement of the hydroxyl group (quantitated by 1H NMR). In contrast, Soine and co-workers [15] describe chlorination of each alkaloid resulting in a mixture of ()-chloroephedrine. Allen et al. state that while 1H NMR analysis of the chloroalkaloid product demonstrate that the product is pure, analysis by GCMS indicates the presence of cis- and/or trans-1,2dimethyl-3-phenylaziridine. The aziridine compounds are presumably produced by intramolecular ring closure, as the chloro-derivative is introduced into the high temperature conditions of the GCMS. The 1H NMR analysis of the chloro-substituted products derived from ()-ephedrine and (+)-pseudoephedrine were in agreement with the results reported by Allen and Kiser [14]. Our GCMS analysis results of the chloro-alkaloid products were also in agreement with those literature results with the cis- and trans-1,2-dimethyl-3-phenylaziridines observed in

similar ratios (Fig. 8) [14]. In our work, chlorination of ephedrine led to >99% pure chloropseudoephedrine (quantitated by 1H NMR) whereas pseudoephedrine led to an approximately 80:20 mixture of chloro-ephedrines, favouring the SNi product chloropseudoephedrine.4 NMR analysis of the product derived from ()-norephedrine indicated a single compound (>99%), presumably (+)chloronorpseudoephedrine, while the product derived from (+)-norpseudoephedrine appeared to be a mixture of (+)chloronorpseudoephedrine and ()-chloronorephedrine (approximately 80:20 favouring chloronorpseudoephedrine). As close analogues of ()-ephedrine and (+)-pseudoephedrine, it is expected that the chlorine atom substitutions proceed via the same SN2 or combination of SN2 and SNi mechanisms, respectively. The cis- and trans-2-methyl-3-phenylaziridines were not observed during NMR analysis, but were again evident in the GCMS analysis of the chloro-alkaloid products. The aziridines were not well resolved and could not be accurately attributed to the cis- or trans-isomers without further work. In this instance, both chloro-products manifested with a similar cis:trans aziridine ratio (Fig. 9). When ()-methylephedrine was treated with thionyl chloride, it was expected that the single isomer of (+)chloromethylpseudoephedrine would be observed via SN2 substitution, whereas (+)-methylpseudoephedrine should give a mixture of ()-chloromethylephedrine and (+)-chloromethylpseudoephedrine via both SN2 and SNi substitution. It was clear from NMR data obtained, that methylephedrine actually yielded a mixture of chloromethylpseudoephedrine and chloromethylephedrine in an 80:20 mixture. In this instance, it appeared that the SNi substitution was more favourable and some of the product was formed via the aziridine intermediate. As expected methylpseudoephedrine yielded a mixture of chloromethylpseudoephedrine and chloromethylephedrine, although in a 95:5 mixture, thus favouring the SNi mechanism more than the previous analogues. The increased proportion of the stereochemistry retained SNi product in both methylephedrine and methylpseudoephedrine reactions can be attributed to the inductive effect of an additional alkyl group on the nitrogen, stabilizing the quaternary aziridine intermediate. Further evidence that these reactions proceed, in some part, through the SNi pathway is observed in an additional minor byproduct evident in the NMR of each dimethyl-chloroproduct. The new byproduct has been tentatively identified as 1dimethylamino-1-phenyl-2-chloropropane. The aziridine intermediate can undergo favourable C1 attack to furnish the expected SNi product, or C2 attack to yield the less favoured byproduct (Fig. 10). The aziridine intermediates, seen previously, were not observed in the NMR results or during GCMS analysis. Instead the same aryl alkenes that were detected during the ‘‘HI’’

4

Fig. 8. GCMS analysis of chloro-alkaloids derived from (+)-pseudoephedrine (top) and ()-ephedrine (bottom).

While chloropseudoephedrine and chloroephedrine cannot be distinguished by GCMS (same retention time and fragmentation fingerprint), they are easily resolved using NMR spectroscopy.

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

107

Fig. 9. GCMS analysis of chloro-alkaloids derived from (+)-norpseudoephedrine (top and bottom-left) and ()-norephedrine (bottom-right).

reduction of ()-methylephedrine and (+)-methylpseudoephedrine were observed by GCMS. In these cases the alkenes can again be tentatively attributed to 1-propenylbenzene and 2propenylbenzene. The new byproduct (1-dimethylamino-1phenyl-2-chloropropane) was also observed in the GCMS data, eluting slightly earlier than each of the chloro-products (Fig. 11). The palladium catalysed hydrogenation of the chloroalkaloids all proceeded in high yield, with no discernible byproducts. As expected the chloro-intermediates derived from ()-ephedrine and (+)-pseudoephedrine yielded (+)-methamphetamine, those derived from ()-norephedrine and (+)-

Fig. 10. Hypothesised mechanism of formation of product and by-product by SNi substitution of (+)-methylpseudoephedrine.

norpseudoephedrine yielded (+)-amphetamine and those derived from ()-methylephedrine and (+)-methylpseudoephedrine yielded N,N-dimethylamphetamine. The hydrogenation product of 1-dimethylamino-1-phenyl-2-chloropropane produced during the chlorination of methylephedrine and

Fig. 11. GCMS analysis of chloro-alkaloids derived from ()-methylephedrine (top) and mass spectra of compound tentatively identified as 1-dimethylamino1-phenyl-2-chloropropane (bottom).

108

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109

Table 1 Summary of products, byproducts and intermediates produced on reduction of Ephedra derived alkaloids Starting material

Reduction method

Product

Intermediates (Observed by GCMS)

Byproducts (Observed by GCMS)

Ephedrine/ pseudoephedrine

‘‘HI’’

Methamphetamine

cis-/trans-1,2-Dimethyl-3-phenylaziridine

Norephedrine/ norpseudoephedrine

‘‘HI’’

Amphetamine

cis-/trans-3-Phenyl-2-methylaziridine

Methylephedrine/ methylpseudoephedrine Ephedrine/pseudoephedrine

‘‘HI’’

N,N-Dimethylamphetamine Methamphetamine

1-Propenylbenzene 2-propenylbenzene Chloropseudoephedrine chloroephedrine cis-/trans-1,2-dimethyl-3-phenylaziridine Chloronorpseudoephedrine chloronorephedrine cis-/trans-2-methyl-3-phenylaziridine Chloromethylpseudoephedrine chloromethylephedrine 1-dimethylamino-1-phenyl-2chloropropane

1-Phenyl-2-propanone 1,3-dimethyl-2-phenylnaphthalene 1-benzyl-3-methylnaphthalene 1-Phenyl-2-propanone 1,3-dimethyl-2-phenylnaphthalene 1-benzyl-3-methylnaphthalene N-(b-phenylisopropyl)benzylmethylketimine 1-Phenyl-2-propanone

Metal catalysed hydrogenation

Norephedrine/ norpseudoephedrine

Metal catalysed hydrogenation

Amphetamine

Methylephedrine/ methylpseudoephedrine

Metal catalysed hydrogenation

N,N-Dimethylamphetamine

Ephedrine/ pseudoephedrine Norephedrine/ norpseudoephedrine Methylephedrine/ methylpseudoephedrine

Dissolving metal (Li/NH3) Dissolving metal (Li/NH3) Dissolving metal (Li/NH3)

Methamphetamine

1-(10 ,40 -Cyclohexadienyl)-2-methylaminopropane (CAP) 1-(10 ,40 -Cyclohexadienyl) aminopropane (CMP) 1-(10 ,40 -Cyclohexadienyl)-2,2-dimethylaminopropane (CDP)

Amphetamine N,N-Dimethylamphetamine

methylpseudoephedrine was not detected, but may co-elute with the strong N,N-dimethylamphetamine peak during GCMS analysis. 3. Summary Previous studies have identified a number of intermediates and byproducts produced during the manufacture of amphetamines. Further compounds, derived from the reduction of Ephedra based alkaloids by several common methods of amphetamine manufacture have been identified or hypothesised and have been described in this article. The products, byproducts and intermediates from the three common reduction methods investigated in this article are summarised in Table 1. 4. Conclusions A number of intermediates and byproducts produced during the reduction of Ephedra alkaloids using several common methods of clandestine methamphetamine manufacture have been identified or hypothesised. The identification of these intermediates and byproducts by GCMS will aid the forensic chemist, when endeavouring to ascertain the source of precursors used in the manufacture of methamphetamine. The information provided here, not only aids the forensic chemist in identifying Ephedra as a precursor for methamphetamine manufacture, but also assists in the elucidation of the synthetic pathway used during the manufacturing process.

References [1] A. Allen, T.S. Cantrell, Synthetic reductions in clandestine amphetamine and methamphetamine laboratories: a review, Forensic Sci. Int. 42 (3) (1989) 183–199. [2] K. Hutchinson, K.M. Andrews, The use and availability of Ephedra products in the United States, Microgram 28 (8) (1995) 256–263. [3] L. Reti, Ephedra Bases, The Alkaloids: Chemistry and Physiology, vol. 3, 1953, pp. 339–362 (Chapter 23). [4] K.M. Andrews, Ephedra’s role as a precursor in the clandestine manufacture of methamphetamine, J. Forensic Sci. 40 (4) (1995) 551–560. [5] L. Pederson, Methamphetamine synthesized from Ephedra extract encountered, J. Clandestine Lab. Investig. Chem. Assoc. 4 (3) (1994) 16–17. [6] H.F. Skinner, Methamphetamine synthesis via hydriodic acid/red phosphorus reduction of ephedrine, Forensic Sci. Int. 48 (2) (1990) 123– 124. [7] T.S. Cantrell, B. John, L. Johnson, A.C. Allen, A study of impurities found in methamphetamine synthesized from ephedrine, Forensic Sci. Int. 39 (1) (1988) 39–53. [8] K.L. Windahl, M.J. McTigue, J.R. Pearson, S.J. Pratt, J.E. Rowe, E.M. Sear, Investigation of the impurities found in methamphetamine synthesised from pseudoephedrine by reduction with hydriodic acid and red phosphorus, Forensic Sci. Int. 76 (1995) 97–114. [9] L. Bernardi, B. Bonini, M. Comes-Franchini, M. Fochi, G. Mazzanti, A. Ricci, G. Varchi, Synthesis and reactivities of enantiomerically pure bhydroxyalkyl and b-aminoalkyl ferrocenyl sulfides, Eur. J. Org. Chem. 26 (2000) 2776–2784. [10] P. Vallely, A single step process for methamphetamine manufacture using hypophosphorus acid, J. Clandestine Lab. Investig. Chem. Assoc. 5 (2) (1995) 14–15. [11] R.A. Ely, D.C. McGrath, Lithium–ammonia reduction of ephedrine to methamphetamine: an unusual clandestine synthesis, J. Forensic Sci. 35 (3) (1990) 720–723.

W.D. Barker, U. Antia / Forensic Science International 166 (2007) 102–109 [12] E.C. Person, J.A. Meyer, J.R. Vyvyan, Structural determination of the principal byproduct of the lithium-ammonia reduction method of methamphetamine manufacture, J. Forensic Sci. 50 (1) (2005) 1–9. [13] T. Dal Cason, A Review of the Birch Reduction Method, Clandestine Laboratory Investigating Chemists Association Monograph, 1998. [14] A.C. Allen, W.O. Kiser, Methamphetamine from ephedrine: 1. Chloroephedrines and aziridines, J. Forensic Sci. 32 (4) (1987) 953– 962. [15] V. Lekskulchai, K. Carter, A. Poklis, W. Soine, GC–MS analysis of methamphetamine impurities: reactivity of (+)- or ()-chloroephedrine

[16] [17]

[18]

[19]

109

and cis- or trans-1,2-dimethyl-3-phenylaziridine, J. Anal. Toxicol. 24 (2000) 602–605. A.J. Birch, Reduction by dissolving metals. Part 1, J. Chem. Soc. (1944) 430–436. S.S. Hall, S.D. Lipsky, G.H. Small, Selective lithium–ammonia reduction of aromatic ketones and benzyl alcohols: mechanistic implications, Tetrahedron Lett. 12 (1971) 1853–1854. G. Zvilichovsky, I. Gbara-Haj-Yahia, Birch reduction of ()-ephedrine. Formation of a new versatile intermediate for organic synthesis, J. Org. Chem. 69 (16) (2004) 5490–5493. H. Emde, Uber Diastereomerie III. Chlor- und brom-ephedrine, Helvetica Chemica Acta 12 (1929) 384–399.

Forensic Science International 166 (2007) 110–114 www.elsevier.com/locate/forsciint

Examination of a long-term clozapine administration by high resolution segmental hair analysis Detlef Thieme a,b,*, Hans Sachs a,b a

Institute of Forensic Medicine, Frauenlobstr. 7a, 80337 Munich, Germany b Forensic Toxicological Centre, Bayerstr. 53, 80335 Munich, Germany

Received 3 December 2005; received in revised form 17 April 2006; accepted 21 April 2006 Available online 12 June 2006

Abstract The long-term administration of clozapine could be verified by fine segmentation and analysis of single hairs of one person to examine the history of a multiple poisoning case. Segments of 1–2.5 mm length were extracted by ultrasonification in 30 ml of the mobile phase (mixture of methanol + water, 50 + 50). By application of isocratic liquid chromatography and using narrow bore columns (Synergy Polar-RP, Phenomenex), an acceleration and miniaturization of the HPLC–MS–MS assay could be achieved. Total amounts of clozapine down to 30 fg (on column) and its desmethyl metabolite could be analysed in multiple reaction monitoring mode. According to typical sample amounts of 16 mg, relevant hair concentrations higher than 1 pg/mg were detected. Significant and reproducible concentration profiles along the hair fibres revealed characteristic administration cycles. The administration time course -in particular the time of its termination—could be verified with a precision of a few days. The accuracy and reproducibility of the concentration profile was proven based on multiple investigations of single hairs. An individual hair growth rate of 0.55 mm/day was determined with a relative standard deviation of 8% by comparison of concentration profiles in hairs collected after a time span of 165 days. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Liquid chromatography–tandem mass spectrometry; Drug monitoring; Hair analysis; Segmentation; Antipsychotic drugs; Clozapine

1. Introduction The identification of drugs of abuse in hair is a wellestablished routine procedure in workplace testing, driving licensing, clinical compliance control and a number of forensic issues (e.g. drugs abstinence control, evaluation of drug induced diminished responsibility and drug facilitated crimes). On the other hand, the lack of correlation between dosage and hair concentration, influences of hair colour and treatment, a restricted reproducibility of analyses caused by heterogeneous specimens and the potential impact of external contaminations are undisputed limitations of this technique. The suitability of hair analysis to the identification of pharmaceutical substances depends on its chemical properties. Basic and lipophilic drugs (e.g. cocaine, clenbuterol) are well incorporated into hair and less susceptible to wash-out effects.

* Tel.: +49 89 54308 135; fax: +49 89 54308 134. E-mail address: [email protected] (D. Thieme). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.015

Clozapine in particular is well suitable for hair analysis [1]. The detection of the parent substance in concentrations up to 59.2 ng/mg [2] and a potential confirmation by its desmethyl metabolite [3,4] were reported. Hair colour appears to be an important parameter for the incorporation rate of clozapine into hair, indicating a high melanin binding of the drug. Moreover, a comparatively good correlation between dosage and hair concentration was reported [5]. Hair segmentation to elucidate the detailed long-term drug history was primarily focused on therapy control of psychiatric patients, e.g. detection of amitriptyline [6], antidepressants and antipsychotic drugs [2], benzodiazepines [7], chlorpromazine [8], clozapine[4], or haloperidol [9,10], other attempts were devoted to the identification of chloroquine [11], piritramide [12] and selegeline compliance control by identification of its metabolites [13]. The use of hair segmentation to receive detailed kinetic information is restricted by substance concentration in hair and corresponding detection limits. Typical hair concentrations range from 1 pg/mg to several ng/mg, involving substance

D. Thieme, H. Sachs / Forensic Science International 166 (2007) 110–114

amounts of 20–100 mg hair. Typical segment lengths are greater than 1 cm of an aligned hair strand, representing an average drug administration time window of 1 month. In a case of the occurrence of characteristic health problems (dizziness, somnolence, unconsciousness, stomach-ache, swollen tongue, speech disorders, apoplectic stroke, circulatory failure, diarrhoea, constipation [14]) amongst at least 46 employees of a service company, an unauthorized long-term administration of clozapine was suspected. Two main episodes of apparent intoxications symptoms, medical consultations and hospitalisations were observed during a time period of 1 year. Therefore, hair testing was undertaken to collect retrospective evidence of poisoning. Clozapine was detected in 24 of the hair samples at concentrations up to 1400 pg/mg. The main goal of the fine segmentation of hair was the accurate differentiation of individual intervals of drug administration and the elucidation of its termination, which was of great importance for the legal assessment of the case. The use of single hairs proved to be essential to overcome statistical uncertainties resulting from unequal growth rates and alignment of individual hairs in hair bundles. 2. Experimental 2.1. Hair sample preparation The following data were obtained from black hair samples of a 35 years old female victim (Asian ethnicity). In the selected case, a second sample was taken after a time span of 165 days. Hair samples were originally collected according to the standard procedure for drug testing, i.e. by cutting a hair strand close to the root and subsequent fixation. Typical initial weights for one analysis were in the order of magnitude 50 mg hair. Hair samples were decontaminated by 5 min agitation in a gas-tight tube with 5 mL petroleum benzene (boiling range to 40 8C, Merck), dried and cut into pieces of 1–2 mm length. After adding 100 ng of the internal standard (MPPH, 5-(4methylphenyl)-5-phenyl hydantoine, research grade, Serva), the hair particles were extracted by 3 h ultrasonification at 55 8C with 3 mL of methanol (Merck, for chromatography). A volume of 1.5 mL of this extract was evaporated and reconstituted with 100 mL of mobile phase. The initial hair segmentations were carried out using 3 cm segments of the hair strand. Afterwards, the amount of hair was reduced to individual hairs, segmented into pieces of 2.5 and 1 mm, respectively. The segmentation of individual hairs required a fixation of the fibre using adhesive tape. Due to the low total amount of hair sample, the mass of single snippets could not be

111

weighted by conventional laboratory devices. Hair concentrations were obtained by converting the length of a segment using an average mass of 6.4 mg/mm (estimated for the respective hair sample). The resulting hair particles are transferred into vials, containing 5 ng of the internal standard (MPPH) in 30 ml of a mixture of water and methanol (50/50, v/v). After three hours of ultrasonification at 55 8C, the vials are acclimatised to ambient temperature and analysed by LC–MS. 2.2. Instrumental analysis All analyses were carried out using an Agilent 1100 LC system (binary pump and autosampler) coupled to an API 4000 mass spectrometer (Applied Biosystems), equipped with a Turbo-Ion-Spray (ESI) source. The optimum ionisation of clozapine (free base, Sandoz) was achieved in positive mode using the parameters described in Table 1. Owing to the high number of individual samples, a rapid chromatographic separation was required. The application of a Synergy Polar-RP (Phenomenex, 75 mm  2.0 mm, 4 mm particle size) column provided a sufficient retention of all analytes under isocratic conditions at ambient temperature. The mobile phase was a 2 mM ammonium acetate buffer (ACS certified, Baker) in a mixture (50/50, v/v) of water (for chromatography, Merck) and acetonitrile (gradient grade, Baker). The mobile phase flow rate was set to 700 mL/min, compatible with a source temperature of 650 8C and source gas flow settings (nitrogen as sprayer and heater gas) of 50 psi. The injection volume was 10 mL. 2.3. Quantitation The accuracy of the quantitative results is mainly governed by parameters other than the conventional uncertainty of analytical processes. Two major factors contribute to the potential quantitative variance: the unknown recovery from hair matrix and the uncertainty of the exact amount (length) of single hair segments. The former represents a systematic error, which is likely to be equal for each individual sample and does not significantly influence concentration profiles or metabolite ratios. The masses of individual pieces were estimated by division the weight of an intact hair by the number of apparently equal segments. The significant inaccuracy of the fine segmenting, possible inhomogeneities of individual hairs and the lacking possibility to control the weight of hair segments results in an estimated statistical error of 20% (referred to the most frequently used 2.5 mm segments). This is supposed to override all other analytical and technical uncertainties by far.

Table 1 Optimum ionisation and fragmentation conditions for clozapine, norclozapine and MPPH

Clozapine Clozapine (qualifier) Norclozapine MPPH

Precursor ion (Da)

Product ion (Da)

Declustering potential (V)

Collision energy (V)

327.3 327.3 313.2 267.1

270.2 192.1 270.2 196.1

91 91 91 71

31 59 31 29

112

D. Thieme, H. Sachs / Forensic Science International 166 (2007) 110–114

Fig. 2. Conventional clozapine concentration profile, obtained from a hair strand cut into 3–5 cm pieces (50 mg each). A concentration maximum is observed in the 6–9 cm segment of the hair strand.

Fig. 1. Fast chromatographic separation of clozapine (RT = 0.64), norclozapine (RT = 0.55) and MPPH (int. std. RT = 0.69, not shown) in a 2.5 mm segment of a hair sample (total weight approximately 16 mg). Comparison of a positive segment (concentration of clozapine 17 pg/mg, equivalent to 90 fg on column, S/N > 24, lower window) and a blank, corresponding to samples ‘1’ and ‘0’ in Fig. 6.

Fig. 3. Comparison of clozapine concentration profiles obtained from two individual hairs collected 165 days after the first hair strand (compare Figs. 4 and 5). Both profiles are in good temporal coincidence, while absolute concentrations differ significantly.

The calibration was performed in a range of 1–5 pg (total amount of clozapine per segment). An appropriate linearity of the calibration is indicated by a regression coefficient better than 0.996. The semi-quantitative extrapolation ranged from 0.1 pg (corresponding to a signal-to-noise ratio of 11) to a maximum of 37 pg (at an intensity of 2  104 cps, positively within the linear detector range). The chromatographic performance is demonstrated in Fig. 1 by comparison of a positive hair segment with a blank segment (sample identity as denoted in Fig. 6). 3. Results and discussion 3.1. Concentration profiles The examination of 3–5 cm pieces of a hair strand indicates only a vague concentration maximum around 9 cm (Fig. 2). The comparison with analytical results, obtained from 2.5 mm segments of single hairs (Figs. 3–5) reveals some distinct and reproducible particularities:  There are at least five different sharp concentration maxima, confirmed by a total number of seven individual hair profiles.

Fig. 4. Comparison of clozapine concentration profiles obtained from 2.5 mm segments of four individual hairs. All concentration profiles show a characteristic pattern of concentration maxima. The peak concentrations are influenced by the random collection of the segments, uncertainty of total sample amount and local hair characteristics.

D. Thieme, H. Sachs / Forensic Science International 166 (2007) 110–114

Fig. 5. The reduction of segment sizes to 1 mm (corresponding to a time period of approximately 2 days) reveals a clear bimodal shape of the first peak and is in good accordance to the data obtained from 2.5 mm segments (compare segment number 1–12, Fig. 4).

 The first (i.e. closest to the root) maximum exhibits a characteristic peak shape of a maximum modified by leading and tailing shoulders.  Local peak concentrations of clozapine in the respective segments are accordingly higher than averages obtained from longer segments.  The location (distance from root) of the concentration maxima shows a very good reproducibility, while absolute concentrations ratios between the peaks may vary significantly. The first concentration maximum was shifted in the 165 days between both sample collections by 9.1 cm. This indicates a relatively high individual growth rate of 0.55 mm/day (1.65 cm/month). Moreover, the variation of growth between individual hairs appears to be unexpectedly low. Estimation of the distance between the first and the fifth concentration maximum of each of the individual hairs results in an average of 10.1 cm and a standard deviation of 7.2%. In contrast to this high reproducibility of the qualitative profile, there are some obvious and significant deviations of quantitative results. A comparison of corresponding segments of different hairs may well exhibit different concentration ratios, which is most likely due to variations of an individual hair structures (e.g. melanin granules). These quantitative variations and the heterogeneity of hairs within a strand (i.e. the presence of 1% of non-growing telogenic hair) requires a repetition of hair profiling for confirmation. The concentration profile of clozapine is paralleled by the corresponding time course of its metabolite norclozapine. The overlay of both time courses (normalised to the respective maximum values, Fig. 6) shows a good match of both curves. This is in good accordance to the expectations, because the resolution of 2.5 mm is equivalent to a time between adjacent segments of 5 days and is therefore larger by far than the plasma half life of clozapine (14 h, 5–60 h) [14]).

113

Fig. 6. Overlay of clozapine and norclozapine (normalised to clozapine) concentration profiles. The results are based on a 2.5 mm segmentation of a single hair from the second sampling. There is no obvious time shift between both concentration profiles, which is in good accordance to the short half life of clozapine. The labels ‘0’ and ‘1’ signify the hair segments corresponding to the chromatographic data as shown in Fig. 1.

3.2. Resolution Hair segment lengths were examined in a range between 1 mm and 3 cm length showing consistent qualitative results. The optimum size of a segment depends on the concentration and detection limits of the target analytes. If the analytical sensitivity is sufficiently high, a segment size of 3–5 mm appears to be an adequate compromise between resolution and reproducibility. Smaller segments may be useful, if the time of a single administration is more relevant than quantitative consistency between adjacent segments. 3.3. Relevance of contamination and diffusion Examination of the concentration alteration at the descending sections of the concentration profile (i.e. after termination of drug administration, e.g. first four segments in Fig. 4), suggests an average reduction of the hair concentration in adjacent 1mm segments by a factor of 3–4. This corresponds to an estimated incorporation half life of the drug in hair of 0.8 days, which is in good agreement to the elimination half life of clozapine in blood. This suggests that the incorporation of clozapine is closely related to blood concentration and it is not likely to be induced by sebum or sweat. Moreover, a potential influence of external contaminations or significant amounts of diffusion of clozapine along the concentration gradient can be excluded, due to the occurrence of sharp and reproducible concentration maxima. 4. Conclusion The diagnostic value of segmental analyses of hair samples for investigations of the drug history is undisputed, e.g. to investigate drug history [15] or treatment compliance in patients [7,13]. However, the accuracy of this information is so far restricted by the detection limits, requiring relatively large

114

D. Thieme, H. Sachs / Forensic Science International 166 (2007) 110–114

amounts of hair. Moreover, the use of bundles of hair generates random uncertainties due to the misalignment of individual hairs. High resolution single hair segmentation may reveal accurate information on drug administration kinetic. Segments of single hair may be downsized to 1 mm length if hair concentration and detection limits provide so. The resulting concentration profiles proved to be highly reproducible and a time resolution of a few days may be achieved. The appearance of sharp concentration maxima in the clozapine hair profiles demonstrates that external contamination and diffusion processes do not lead to concentration alterations. A second sampling after 165 days showed consistent time course. An individual growth rate of 1.65 cm/month may be estimated from superposition of both concentration profiles. The generalisation of this approach is certainly restricted, attempts to profile clozapine and other antipsychotic drugs in hair of other individuals showed a potential influence of external contamination and/or wash-out effects. In general, the accomplishment of the procedure depends on the chemical properties of the compound in combination with its specific hair incorporation rate and respective results need a careful and case related interpretation.

[4] [5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

References [1] B. Ahrens, G. Rochholz, H.W. Sachs, H. Schutz, Detection of clozapine in hair after 1 years burial in soil grave, Arch Kriminol 196 (1995) 138. [2] M. Shen, P. Xiang, H. Wu, B. Shen, Z. Huang, Detection of antidepressant and antipsychotic drugs in human hair, Forensic Sci Int 126 (2002) 153. [3] W. Weinmann, C. Muller, S. Vogt, A. Frei, LC–MS–MS analysis of the neuroleptics clozapine, flupentixol, haloperidol, penfluridol, thioridazine,

[14]

[15]

and zuclopenthixol in hair obtained from psychiatric patients, J Anal Toxicol 26 (2002) 303. M. Rothe, Math.-Nat Fakulta¨t, Humboldt Universita¨t, Berlin, 1997. V. Cirimele, P. Kintz, O. Gosselin, B. Ludes, Clozapine dose-concentration relationships in plasma, hair and sweat specimens of schizophrenic patients, Forensic Sci Int 107 (2000) 289. F. Pragst, M. Rothe, J. Hunger, S. Thor, Structural and concentration effects on the deposition of tricyclic antidepressants in human hair, Forensic Sci Int 84 (1997) 225. R. Kronstrand, I. Nystrom, M. Josefsson, S. Hodgins, Segmental ion spray LC–MS–MS analysis of benzodiazepines in hair of psychiatric patients, J Anal Toxicol 26 (2002) 479. H. Sato, T. Uematsu, K. Yamada, M. Nakashima, Chlorpromazine in human scalp hair as an index of dosage history: comparison with simultaneously measured haloperidol, Eur J Clin Pharmacol 44 (1993) 439. M. Nakano, T. Uematsu, H. Sato, K. Kosuge, M. Nishimoto, M. Nakashima, Using ofloxacin as a time marker in hair analysis for monitoring the dosage history of haloperidol, Eur J Clin Pharmacol 47 (1994) 195. T. Uematsu, H. Matsuno, H. Sato, H. Hirayama, K. Hasegawa, M. Nakashima, Steady-state pharmacokinetics of haloperidol and reduced haloperidol in schizophrenic patients: analysis of factors determining their concentrations in hair, J Pharm Sci 81 (1992) 1008. U. Runne, F.R. Ochsendorf, K. Schmidt, H.W. Raudonat, Sequential concentration of chloroquine in human hair correlates with ingested dose and duration of therapy, Acta Derm Venereol 72 (1992) 355. D. Thieme, H. Sachs, Progress in forensic toxicology by application of liquid chromatography–mass spectrometry, Anal Chim Acta 492 (2002) 171. R. Kronstrand, J. Ahlner, N. Dizdar, G. Larson, Quantitative analysis of desmethylselegiline, methamphetamine, and amphetamine in hair and plasma from Parkinson patients on long-term selegiline medication, J Anal Toxicol 27 (2003) 135. R.J. Flanagan, E.P. Spencer, P.E. Morgan, T.R.E. Barnes, L. Dunk, Suspected clozapine poisening in the UK/Eire, 1992–2003, Forensic Sci Int 155 (2004) 91–99. K.M. Clauwaert, J.F. Van Bocxlaer, W.E. Lambert, A.P. De Leenheer, Segmental analysis for cocaine and metabolites by HPLC in hair of suspected drug overdose cases, Forensic Sci Int 110 (2000) 157.

Forensic Science International 166 (2007) 115–120 www.elsevier.com/locate/forsciint

A distinct Y-STR haplotype for Amelogenin negative males characterized by a large Yp11.2 (DYS458-MSY1-AMEL-Y) deletion Yuet Meng Chang a,*, Revathi Perumal a, Phoon Yoong Keat a, Rita Y.Y. Yong b, Daniel L.C. Kuehn c, Leigh Burgoyne c b

a Forensic DNA Laboratory, Department of Chemistry, Petaling Jaya, Malaysia Defence Medical and Environmental Research Institute, DSO National Laboratories, Singapore, Singapore c School of Biological Sciences, Flinders University, Adelaide, Australia

Received 10 September 2005; received in revised form 16 April 2006; accepted 21 April 2006 Available online 9 June 2006

Abstract The use of STR multiplexes with the incorporated gender marker Amelogenin is common practice in forensic DNA analysis. However, when a known male sample shows a dropout of the Amelogenin Y-allele, the STR system falsely genotypes it as a female. To date, our laboratory has observed 18 such cases: 12 from our Y-STR database and six from casework. A study on 980 male individuals in the Malaysian population using the AmpFlSTR1 Y-filerTM has revealed a distinct Y-chromosome haplotype associated with the Amelogenin nulls. Our results showed that whilst the Amelogenin nulls were noticeably absent among the Chinese, both the Indians and Malays exhibited such mutations at 3.2 and 0.6%, respectively. It was also found that the Amelogenin negative individuals predominantly belonged to the J2e lineage, suggesting the possibility of a common ancestor for at least some of these chromosomes. The null frequencies showed concordance with the data published in Chang et al. (Higher failures of Amelogenin sex test in an Indian population group, J. Forensic Sci. 48 (2003) 1309–1313) [1] on a smaller Malaysian population of 338 males which used a Y-STR triplex. In the current study, apart from the absence of the Amelogenin Y-locus, a complete absence of the DYS458 locus in all the nulls was also observed. This study together with the 2003 study has indicated a similar deletion region exists on the Yp11.2 band in all the 18 Ychromosomes. Using bioinformatics, this deletion has been mapped to a region of at least 1.13 Mb on the Yp11.2 encompassing the Amelogenin, MSY1 minisatellite and DYS458 locus. Further, the Y-filerTM haplotypes revealed an additional null at Y-GATA H4 in two of the Indian males presented here. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Amelogenin (AMEL); Amelogenin Y-allele (AMEL-Y); Null allele; Haplotype; Deletion; Indians; Malays; Chinese; Malaysia

1. Introduction In forensic science, gender is commonly determined through PCR-based assays that target a small region of the X–Y homologous Amelogenin gene. This sex-typing test is easily performed using the AMEL 106/112 bp primers that are routinely incorporated in standard forensic STR multiplexes and generates two amplicons of 6 bp difference, the longer product being the Ychromosome amplicon. Dropout of the AMEL-X (106 bp) has been shown to be due to a mutation in the primer-binding region of the gene on the X-chromosome [2], though this will not cause gender mistyping as the Y-allele is still present. However, if the * Corresponding author. Tel.: +60 3 79853836; fax: +60 3 79581173. E-mail addresses: [email protected], [email protected] (Y.M. Chang). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.013

AMEL-Y has dropped out, the STR system will falsely genotype the sample as a female. The Forensic DNA Laboratory of the Department of Chemistry in Malaysia has observed a number of such cases from the Malaysian population. The occurrence of AMEL-Y negative males in the Malaysian population was first reported in a study on 113 Malays, 113 Chinese and 112 Indians using a 3-loci Y-STR multiplex in 2003 [1]. Recently, we carried out Y-STR haplotyping on a larger male population comprising of 334 Malays, 331 Chinese and 315 Indians using a 16-loci multiplex, the Y-filerTM by Applied Biosystems (Foster City, CA, USA). This megaplex simultaneously amplifies nine European minimal loci (DYS19, DYS389I, DYS389II, DYS390, DYS391, DYS392, DYS393, DYS385a/b), two SWGDAM-extended loci (DYS438, DYS439) and six new additional loci (DYS456, DYS458, DYS635 or YGATA C4, Y-GATA H4, DYS437 and DYS448).

116

Y.M. Chang et al. / Forensic Science International 166 (2007) 115–120

To date, our laboratory has observed 18 AMEL-Y negative males: 12 from our Y-STR database (10 Indians, 2 Malays) and 6 from casework (5 Indians, 1 Malay). We describe here discordant observations where the AMEL Y-locus, along with the DYS458 locus, was completely absent in the 18 Ychromosomes. In all these cases, we report a failure in the sex genotyping of the male individuals. Except for four null males, the haplogroup or haplogroups where the deletion took place were also ascertained. Based on the results obtained, a brief discussion of the plausible origins of the nulls is included.

Integrated System (IS) software of the Liquidchip workstation, and data analysis with Microsoft1 Excel. A two-level screen was employed: the first screen classified samples into major clades following the YCC 2003 phylogenetic tree [3], whilst the second screen delineated specific haplogroups within each clade:

2. Materials and methods

 J2e was identified by its Y-SNP signature of M89T/M9C/ M304C/M172G/M241A [4];  D* by M168T/M89C/M96G/M130C/M145A/M15(D)/ M55T;  F* by M89T/M9C/M304A/M201G/M52A/M170A/M253C/ P37T/M26G.

2.1. Collection and isolation of genomic DNA

3. Results

Blood of 980 unrelated males with self-reported ethnicity from the three main ethnic populations in Malaysia was stained on FTA1 cards (Whatman1, USA) and a 1.2 mm disc was excised from each card and purified in situ using standard protocols. For the casework samples, Chelex extraction was carried out followed by quantification using QuantifilerTM Real-Time PCR (Applied Biosystems, Foster City, CA, USA).

Based on the Identifiler1 STR results, 18 male individuals exhibited a female genotype, indicated by the absence of the 112 bp AMEL-Y peak. These subjects were coded N1–N18, with the first six described previously [1]. Of the 18 AMEL-Y negative males, 14 were from the Indian group whilst the remaining 4 were from the Malay group. Interestingly, no discordant results were detected in the Chinese group. Subsequent amplification of the 16 Y-filerTM Y-STR loci provided more definitive male haplotypes (Table 1), confirming that the 18 discordant samples were from true male individuals. Preliminary haplogrouping (Hg) studies using Y-SNPs on 14 of the AMEL-Y nulls showed that 12 belong to Hg J2 (N1– N4, N7, N8, N10–N15), with one Hg D (N5) and one Hg H (N9) (personal communication from M.A. Jobling, 2006, see Table 1). The haplogroups for the other four nulls from casework (N6, N16–N18) could not be determined due to insufficient quantities. Recently, 12 of the population database null samples were tested for haplogroups using a different panel of Y-SNP markers. The results showed that 10 out of the 12 nulls (N1–N4, N7, N8, N10–N13) fall under the same Y-chromosome haplogroup which is J2e, with the other two nulls under Hg F* (N9), and Hg D* (N5) (see Table 1). It is interesting to note that the null sample N9 has been typed as H in the first system but F* in the second system. The difference in the haplogroups obtained for N9 is likely due to the M52 marker, which is the critical Y-SNP marker that differentiates Hg H from other haplogroups. The ancestral state of M52 is A, whilst the derived state that defines H is C. Both H and F* samples have the common signature of M89T/ M9C/M304A. For finer differentiation, H will have M52C/ M170A/M201G, whilst F* will have M52A/M170A/M201G. The Malaysian null N9 was confirmed to have M52A as standards from NIST (SRM2395, National Institute of Standards and Technology, USA) with known M52A status were analysed concurrently during the haplogrouping analysis. In addition, there are samples from the Singaporean database that display the alternative status, M52C (data not shown). Therefore the technical performance of M52A and M52C has been validated in the present Y-SNP assay system.

2.2. Y-STR haplotyping Amplification with the Y-filerTM multiplex was performed directly on the purified discs using 20 ml reaction volume following recommendations by the manufacturer on a GeneAmp1 PCR System 9700 (Applied Biosystems). For FTA1 samples, a single purified disc was used in the PCR with reduced (27) cycles. For Chelex extracts, approximately 1– 1.2 ng per assay was used. The amplified products were electrophoretically separated on an ABI Prism1 3100 Genetic Analyzer (Applied Biosystems) using an injection time of 10 s at 3 kV; whilst POP-4TM, loading mix and other reagents were according to the manufacturer’s instructions. Following separation, the fluorescent fragments were analyzed using the Genescan1 and Genotyper1 programs (Applied Biosystems). Amplification with the AmpFlSTR1 Identifiler1 STR system (Applied Biosystems, Warrington, UK) was also carried out for the null samples following recommendations by the manufacturer prior to Y-STR haplotyping. 2.3. Y-SNP haplogrouping Y-SNP haplogrouping was carried out with a beta test version Y-SNP kit manufactured by Marligen Biosciences (Maryland, USA). A microbead platform was employed. YSNP fragments amplified by multiplex PCR were labelled with a fluorescent tag. The labelled products were hybridized to YSNP-specific oligonucleotides immobilised onto individually colour-coded beads. The bead mixes were analysed on a Liquidchip workstation (Qiagen, Germany), which delineated the spectral address of each bead and the amount of fluorescent tag per bead. Fluorescent data acquisition was done with the

Table 1 Y-STR and haplogroup results of the 18 AMEL-Y negative males typed with Y-filerTM

Y.M. Chang et al. / Forensic Science International 166 (2007) 115–120 117

118

Y.M. Chang et al. / Forensic Science International 166 (2007) 115–120

4. Discussion 4.1. Amelogenin-Y null frequencies The data presented here was in agreement with the frequency data published in a 2003 study [1] on a smaller Malaysian population using a Y-STR triplex. However, this larger study has provided more information although still not complete, on the nulls and their possible origin/divergence. The AMEL-Y negative males were mostly from one migrant group which is the Indian group (3.2%), with a small proportion in the native Malay group (0.6%) and noticeably none in the Chinese group, the other major migrant group. A higher AMEL-Y dropout was similarly noted in several global population groups [1,5–11, and personal communication from M.A. Jobling (2005)]. Overall, there is a notably higher frequency of this mutation in males of Indian or Sri Lankan origin and a much lower frequency in a few Caucasian population groups (Table 2). It is now evident that there is a prominent lineage associated with the AMEL-Y null that may have spread from the Indian subcontinent, with other similar mutations cropping up sporadically elsewhere in the world. The conspicuous absence of the AMEL-Y null among the Chinese in Malaysia and in other reported global populations thus far poses an interesting question: how is the AMEL-Y null propagating through a population? What is so different about the Indian and Sri Lankan populations that they have much higher frequencies of AMEL-Y nulls than other populations?

mutation. In contrast, the Y-specific minisatellite MSY1 locus, located on the same Yp11.2 band as the AMEL-Y locus, has been shown to be absent in six previously studied deletion males [1]. Recently, Lattanzi et al. [10] used deletion mapping to locate the distal break-point for this interstitial deletion that encompasses the AMEL-Y locus and confirmed that the lesion spanned approximately 2.5 Mb on the pericentromeric region of the short arm of the Y-chromosome. The Y-filerTM haplotypes associated with this AMEL-Y allele deletion mutation in the 18 males consistently showed an absence of the DYS458 locus, indicating that both loci were similarly deleted. DYS458, a polymorphic complex tetranucleotide (GAAA) repeat locus has seven alleles in the Y-filerTM allelic ladder ranging from 14 to 20, with allele 16 being the most common allele observed in both the Indian (27.0%) and Malay (32.1%) groups in Malaysia [12]. Using published primer sequences [13–16] and the Ensembl human genome browser (http://www.ensembl.org/Homo_sapiens/index.html), the map locations of the 16 Y-filerTM loci, the AMEL-Y, MSY1, SRY and DYZ1 were plotted. DYS458 and the AMEL-Y loci differed by 1.13 Mb on the short arm of the Y-chromosome, with the minisatellite MSY1 located in between the two loci, and 0.73 Mb from the AMEL-Y locus. This indicated a large deletion in the Yp11.2 band in all the affected 18 Y-chromosomes which includes Amelogenin, the MSY1 minisatellite and the DYS458 locus, and are all within the boundaries of the deletion mapped by Lattanzi et al. [10] as shown in the schematic map (Fig. 1). 4.3. Possible origins of the AMEL-Y null

4.2. A limited Yp11.2 deletion (DYS458-MSY1-AMEL-Y) on the 18 Y-chromosomes Previous studies using alternative sets of AMEL primers targeting different regions of the gene have demonstrated that the failure of the Amelogenin sexing test is commonly due to a large lesion encompassing the AMEL copy on the Y-chromosome [1,6,10]. Other suggested sex-markers on the Y-chromosome which are located outside the AMEL gene have also been tested, such as the SRY gene (located on Yp11.31) [5,8–10], and the DYZ1 locus (located on Yq12) [6], and are not commonly involved as amplifications using the primers for these two loci demonstrated that they were intact in males carrying this deletion Table 2 Summary of AMEL-Y null frequencies in global population groups Population

No. of nulls/ individuals studied

Frequency (%)

Reference

Sri Lanka Austria India (general) Italy Israel South India England Spain Malaysian Indians Malaysian Malays

2/24 5/28182 5/270 1/13000 1/96 1/100 2/2000 1/1000 10/315 2/334

8.3 0.018 1.9 0.008 1.0 1.0 0.1 0.1 3.2 0.6

[5] [8] [9] [10] [11] Pers. comm. Pers. comm. Pers. comm. Current study Current study

Note: Pers. comm.: personal communication from M.A. Jobling (2005).

There are generally three primary mechanisms for any mutation to transmit in the human genome: (1) founder effect/ genetic drift, (2) genomic processes (normal mutations) and (3) selection forces (at the level of the individual, or physical and/ or social and/or other bases). It seems unlikely that the frequency of the deletion has arisen as a consequence of natural selection on these Ychromosomes, as there are no distinct classes of related Y-STR haplotypes observed to go with the deletion on Yp11.2 band in the two affected ethnic populations harbouring this deletion (see Table 1). All 18 males’ deletion consistently included the same three null loci of AMEL-Y-MSY1-DYS458 which are all within a region of 1.13 Mb, suggesting that either a single ancestral deletion event involving the three null loci or three random deletion events had occurred on these Y-chromosomes. Y-SNP haplogrouping results revealed that there are at least three different haplogroups in the 18 males carrying the AMELY deletion. Out of the 14 nulls tested for haplogrouping, 12 (10 Indians and 2 Malays) were from haplogroup J2, 10 of these are now tested to belong to sub-clade J2e in the current study. The remaining Indian null belongs to haplogroup F*, which is ancestral to the J2e chromosomes according to the YCC tree [17], with the other Malay null (N5) being from haplogroup D*. The data observed so far indicates a founder event for the J2 deletion some time ago but we do not know the age of this deletion. The other null from a Malay (N5) belonging to haplogroup D* may represent an independent event happening

Y.M. Chang et al. / Forensic Science International 166 (2007) 115–120

119

Fig. 1. Schematic representation of human Y-chromosome map. (A) Approximate locations of each of the 16 Y-STR loci from Y-filerTM in addition to the locations of AMEL-Y, MSY1, SRY, and DYZ1. The locations were determined by performing database searching with published primer sequences. Immediately below the approximate scale are the four regions susceptible to deletion. Region I is the AMEL-Y deletion; Regions II, III, and IV, are the azoospermia factor related regions, AZFa, b, and c, respectively [22–24]. Locations marked with * represent double loci, e.g. DYS385a,b at 19.19 Mb. (B) An expanded view of the AMEL-Y-MSY1DYS458 contigs of the Yp11.2 band with a more accurate scale. The deletion mapped by Lattanzi et al. [10] begins at 6.44 Mb and ends around 9 Mb. Contig accession numbers from the Ensembl Genome browser are included.

sporadically elsewhere but may also involve similar mechanisms to the J2’s chromosomes. The Y-chromosome J2 lineage (defined by mutation M172) is approximately 15,000 years old [4] and expanded throughout the Middle East, Central Asia, the Mediterranean and India about 7,500 years ago [18]. Currently, we do not have the haplogroup results of the larger Indian population in the Malaysian database, however haplogroup J2 is said to exist in approximately 9% of the population in India whilst haplogroup F* (defined by mutation M89) is approximately 5% [19]. An analysis of the average allele repeat number and allele variance observed at certain Y-STR loci (DYS19, DYS389I, DYS390, DYS391, DYS392, DYS393 and DYS439; Table 1) for the J2e nulls in this study corresponds to that reported by Cinnioglu et al. [20] for the J2e haplogroup. It has recently been shown that the absence of Y-STR products from certain loci in the q-arm of the Y-chromosome can inadvertently reveal infertile males in the population [21]. Interestingly, two out of the 14 Indian AMEL-Y nulls also showed nulls at the Y-GATA H4 locus (located on Yq11.221, see Fig. 1), suggesting a more recent deletion around the AZFb

region, which corresponds with one of the three major ‘‘hotspots’’ defined on the long arm of the Y-chromosome responsible for azoospermia (see Table 1 and Fig. 1). Alleles for the Y-GATA H4 locus range from 8 to 13, with allele 12 being the most common allele observed in both the Malay (42.0%) and Indian (51.8%) groups [12]. Due to a lack of genotype– phenotype analysis, it is not known whether the deletions reported here have some compensating advantage or are immune in some way from the expected semen impairment. The 2.5 Mb deletion defined by Lattanzi et al. [10] in the Italian population was found in one infertile oligozoospermic male, one pre-natal male, and their respective fathers. Given that both fathers had produced their sons naturally, they were assumed to be fertile despite the presence of the AMEL-Y deletion, and the infertility of the test subject is therefore presumed to have had other origins [10]. Even though AMEL-Y negative males seem to have multiple origins in global populations, the haplogroup results for the Indian nulls have so far indicated an ancestral J2e lineage, and the presence of an additional null at Y-GATA H4 in two of the affected Indian males may be worth investigating further for

120

Y.M. Chang et al. / Forensic Science International 166 (2007) 115–120

any potential link to the AMEL-Y deletion. Future work should include estimating the age of the deletion mutation with more in-depth mapping, genotype–phenotype analysis or comparison of more haplotypes between the nulls from similar lineages and haplogrouping from the Malaysian population database. Determining the DNA sequence across the deletions would also be potentially valuable to further investigate the actual mutational mechanisms at work. 5. Conclusion The failure rate of the Amelogenin sexing test has a consequence in the forensic scenario in Malaysia and any nation with a sizeable Indian population. Whenever a casework stain shows an AMEL-Y null it cannot be assumed to be femaleorigin, and requires the additional determination of the Y-STR haplotype. The absence of the DYS458 locus and/or the MSY1 minisatellite, in the presence of other Y-STR loci, would then serve as a stronger indication of the AMEL-Y deletion. To the forensic community, although the frequency of the AMEL-Y null is relatively low in population groups other than Indians, gender testing in human identification needs high reliability and therefore a sex genotyping test based on Amelogenin locus alone is insufficient for casework. In this respect, routine YSTR haplotyping using kits like Y-filerTM can provide a quick and practical means to confirm males carrying this chromosomal aberration. Acknowledgements The authors gratefully acknowledge Prof. M.A. Jobling for the personal communications and the two expert reviewers for their helpful comments and suggestions. We would like to thank the Director-General and all staff of the DNA/Serology Laboratory, Forensic Division at the Department of Chemistry, Ministry of Science, Technology and Innovation, Malaysia. We also thank all the volunteers who have kindly donated their blood samples for this study. This study was supported by the Department of Chemistry, Ministry of Science, Technology and Innovation in Malaysia. References [1] Y.M. Chang, L.A. Burgoyne, K. Both, Higher failures of Amelogenin sex test in an Indian population group, J. Forensic Sci. 48 (2003) 1309–1313. [2] B. Shadrach, M. Commane, C. Hren, I. Warshawsky, A rare mutation in the primer binding region of the Amelogenin gene can interfere with gender identification, J. Mol. Diagn. 6 (2004) 401–405. [3] M.A. Jobling, C. Tyler-Smith, The human Y chromosome: an evolutionary marker comes of age, Nat. Rev. Genet. 4 (2003) 598–612. [4] P. Shen, T. Lavi, T. Kivisild, V. Chou, D. Sengun, D. Gefel, I. Shpirer, E. Woolf, J. Hillel, M.W. Feldman, P.J. Oefner, Reconstruction of patrilineages and matrilineages of Samaritans and other Israeli populations from Y-chromosome and mitochondrial DNA sequence variation, Hum. Mutat. 24 (2004) 248–260. [5] F.R. Santos, A. Pandya, C. Tyler-Smith, Reliability of DNA-based sex tests, Nat. Genet. 18 (1998) 103.

[6] P.E. Roffey, C.I. Eckhoff, J.L. Kuhl, A rare mutation in the Amelogenin gene and its potential investigative ramifications, J. Forensic Sci. 45 (2000) 1016–1019. [7] J. Henke, L. Henke, P. Chatthopadhyay, M. Kayser, M. Du¨lmer, S. Cleef, H. Poche, H. Felske-Zech, Application of Y-chromosomal STR haplotypes to forensic genetics, Croat. Med. J. 42 (2001) 292–297. [8] M. Steinlechner, B. Berger, H. Niedersta¨tter, W. Parson, Rare failures in the Amelogenin sex test, Int. J. Legal Med. 116 (2002) 117–120. [9] K. Thangaraj, A.G. Reddy, L. Singh, Is the Amelogenin gene reliable for gender identification in forensic casework and prenatal diagnosis? Int. J. Legal Med. 116 (2002) 121–123. [10] W. Lattanzi, M.C. Di Giacomo, G.M. Lenato, G. Chimienti, G. Voglino, N. Resta, G. Pepe, G. Guanti, A large interstitial deletion encompassing the Amelogenin gene on the short arm of the Y chromosome, Hum. Genet. 116 (2005) 395–401. [11] A. Michael, P. Brauner, Erroneous gender identification by the Amelogenin sex test, J. Forensic Sci. 49 (2004) 258–259. [12] Y.M. Chang, R. Perumal, P.Y. Keat, D.L.C. Kuehn, Haplotype diversity of 16 Y-chromosomal STRs in three main ethnic populations (Malays, Chinese and Indians) in Malaysia, Forensic Sci. Int. [Epub ahead of print]. [13] A. Hall, J. Ballantyne, The development of an 18-locus Y-STR system for forensic casework, Anal. Bioanal. Chem. 376 (2003) 1234–1246. [14] A.J. Redd, A.B. Agellon, V.A. Kearney, V.A. Contreras, T. Karafet, H. Park, P. de Knijff, J.M. Butler, M.F. Hammer, Forensic value of fourteen novel STRs on the human Y-chromosome, Forensic Sci. Int. 130 (2002) 97–111. [15] M.A. Jobling, N. Bouzekri, P.G. Taylor, Hypervariable digital DNA codes for human paternal lineages: MVR-PCR at the Y-specific minisatellite, MSY1 (DYF155S1), Hum. Mol. Genet. 7 (1998) 643–653. [16] A. Akane, H. Shiono, K. Matsubara, Y. Nakahori, S. Seki, S. Nagafuchi, M. Yamada, Y. Nakagome, Sex identification of forensic specimens by polymerase chain reaction (PCR): two alternative methods, Forensic Sci. Int. 49 (1991) 81–88. [17] The Y-chromosome consortium, A nomenclature system for the tree of human Y-chromosomal binary haplogroups, Genome Res. 12 (2002) 339– 348. [18] A. Nebel, D. Filon, B. Brinkmann, P.P. Majumder, M. Faerman, A. Oppenheim, The Y chromosome pool of Jews as part of the genetic landscape of the Middle East, Am. J. Hum. Genet. 69 (2001) 1095–1112. [19] S. Sengupta, L.A. Zhivotovsky, R. King, S.Q. Mehdi, C.A. Edmonds, C.E. Chow, A.A. Lin, M. Mitra, S.K. Sil, A. Ramesh, M.V. Usha Rani, C.M. Thakur, L.L. Cavalli-Sforza, P.P. Majumder, P.A. Underhill, Polarity and temporality of high-resolution Y-chromosome distributions in India identify both indigenous and exogenous expansions and reveal minor genetic influence of Central Asian Pastoralists, Am. J. Hum. Genet. 78 (2006) 202–221. [20] C. Cinnioglu, R. King, T. Kivisild, E. Kalfoglu, S. Atasoy, G.L. Cavalleri, A.S. Lillie, C.C. Roseman, A.A. Lin, K. Prince, P.J. Oefner, P. Shen, O. Semino, L.L. Cavalli-Sforza, P.A. Underhill, Excavating Y-chromosome haplotype strata in Anatolia, Hum. Genet. 114 (2004) 127–148. [21] T.E. King, E. Bosch, S.M. Adams, E.J. Parkin, Z.H. Rosser, M.A. Jobling, Inadvertent diagnosis of male infertility through genealogical DNA testing, J. Med. Genet. 42 (2005) 366–368. [22] F. Raicu, L. Popa, P. Apostol, D. Cimponeriu, L. Dan, E. Ilinca, L.L. Dracea, B. Marinescu, L. Gavrila, Screening for microdeletions in human Y chromosome-AZF candidate genes and male infertility, J. Cell. Mol. Med. 7 (2003) 43–48. [23] P.C. Patsalis, N. Skordis, C. Sismani, L. Kousoulidou, G. Koumbaris, C. Eftychi, G. Stavrides, A. Ioulianos, S. Kitsiou-Tzeli, A. Galla-Voumvouraki, Z. Kosmaidou, C.G. Hadjiathanasiou, K. McElreavey, Identification of high frequency of Y chromosome deletions in patients with sex chromosome mosaicism and correlation with the clinical phenotype and Y-chromosome instability, Am. J. Hum. Genet. 135 (2005) 145–149. [24] G. Vinci, F. Raicu, O. Popa, R. Cocos, K. McElreavey, A deletion of a novel heat shock gene on the Y chromosome associated with azoospermia, Mol. Hum. Reprod. 11 (2005) 295–298.

Forensic Science International 166 (2007) 121–127 www.elsevier.com/locate/forsciint

A new sensitive short pentaplex (ShoP) PCR for typing of degraded DNA C. Meissner a,*, P. Bruse a, E. Mueller c, M. Oehmichen a,b a

Department of Forensic Medicine, Medical University of Luebeck, Kahlhorststrasse 31-35, 23562 Luebeck, Germany b Department of Forensic Medicine, University of Kiel, Germany c Bundeskriminalamt, Wiesbaden, Germany Received 13 May 2005; received in revised form 18 April 2006; accepted 21 April 2006 Available online 30 June 2006

Abstract Analysis of short tandem repeat makers has become the most powerful tool for DNA typing in forensic casework analysis. Unfortunately, typing of DNA extracted from telogen shed hairs, bones buried in the soil or from paraffin-embedded, formalin-fixed tissue often reveals no results due to the degradation of DNA. The reduction in size of the target fragments by development of new primers and their combination in multiplex approaches open a new field of DNA analysis. Here we present a new sensitive short pentaplex PCR including the loci amelogenin, TH01, VWA, D3S1358 and D8S1179. Validation tests of our new method included sensitivity, mixtures, human specificity, artificial degradation of DNA by DNase I and case work analysis on a panel of different forensic samples. The detection limit was 12.5 pg of human DNA, and mixtures of 50 pg in a total of 1000 pg were clearly detectable and revealed complete profiles. Only DNA extracts of human primates displayed a few signals, whereas other animal, fungal or bacterial DNA showed no signals. Our method proved extremely valuable in the analysis of artificially degraded DNA and in forensic cases, where only poorly preserved DNA was available. This approach and other similar methods can aid in the analysis of samples where allelic drop out of larger fragments is observed. It is highly recommended to develop more of these multiplexes to improve poor quality DNA typing. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: DNA typing; Short tandem repeats; Multiplex PCR; Forensic casework

1. Introduction In the last decade, DNA typing of highly polymorphic STR (short tandem repeat) loci has become the most powerful tool for discrimination of individuals, because these loci are even stable in decomposed tissues [1]. In particular, the simultaneous amplification of multiple STR markers in the same PCR has found multiple applications even in cases with only minute amounts of DNA [2]. In these multiplex PCR assays usually fluorescent dyes are attached to the forward primer of different loci to allow discrimination of fragments of similar size [3]. The development of capillary electrophoresis enables the on-line detection of these labelled PCR amplicons with the advantage of high throughput, automatic operation and automated data acquisition [4]. Nevertheless there are a lot of forensic cases

* Corresponding author. Tel.: +49 451 500 2752; fax: +49 451 500 2760. E-mail address: [email protected] (C. Meissner). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.014

where DNA is highly degraded due to the quality of the sample itself [5] or environmental conditions [6,7]. These cases include skeletal remains buried in the soil [8], decomposed bodies [9], paraffin-embedded tissue [10] or shed telogen hairs [5]. The rate of decay of DNA depends to a large extend on the geochemical properties of the soil, the effects of the surrounding milieu, contamination with microorganisms and temperature [7,11,12]. Especially exposure of bone or teeth in damp environments seems to be crucial for successful DNA typing [6,13]. On the other hand, DNA degradation is reduced under permafrost conditions in arctic regions, facilitating analyses of remains up to 50,000 years old [11]. Nevertheless, ancient DNA extracted from bones up to more than 10,000 years old displays an average size between 100 and 150 bp and oxidative as well as hydrolytic damage making PCR amplification and DNA typing extremely difficult [11,14,15]. At worst, DNA can be degraded to such an extent that it is no longer suitable for demonstration of STR profiles [16]. Therefore, a lot of different approaches have been presented to

122

C. Meissner et al. / Forensic Science International 166 (2007) 121–127

improve the quality of DNA profiling of degraded DNA. These studies include the evaluation of different extraction methods [9], removal of special PCR inhibitors, as well as purification of extracts [17,18], a nested PCR method [19] and typing of standardized degraded DNA [12]. Recently an excellent review about the application of reduced size amplicons for reliable DNA typing of degraded DNA has been published. These miniplex PCRs are recommended in cases where allelic drop out and reduced sensitivity especially of larger alleles occurs [20]. So far, a few of these interesting miniplex PCRs have already been described for a variety of forensic samples [20,21] including shed telogen hairs [5]. Learning from lessons of degradation of DNA even over long periods of time we developed a multiplex PCR where none of the detectable alleles is more than 150 bp in length. Here we present a short pentaplex PCR (ShoP-PCR) approach which is highly suitable to obtain excellent results even in cases where DNA typing was unsuccessful using commercially available kits. 2. Materials and methods 2.1. Primer design Sequences of primers were designed using the Primer3 [22] and GeneFisher (http://bibiserv.techfak.uni-bielefeld.de/cgi-in/ gf_submit?mode=STARTUP&sample=dna) software. Forward primers were labelled with fluorescent dyes as shown in Table 1.

incubation at 37 8C for 5 min. The cycling profile of the ShoPPCR in a GeneAmp PCR system 2400 (Applied Biosystems, Darmstadt, Germany) was 95 8C for 11 min (initial incubation), 96 8C for 2 min, followed by 10 cycles of denaturation for 30 s at 94 8C, annealing for 30 s at 60 8C and extension for 45 s at 70 8C and then by 18–22 cycles of denaturation for 30 s at 90 8C, annealing for 30 s at 60 8C and extension for 45 s at 70 8C. This was followed by a final elongation step of 90 min at 60 8C. At the end of the PCR reaction, the temperature was kept at 4 8C. Ramping time between annealing and extension was carefully adjusted between 0.3 and 0.5 8C/s. 2.3. Signal detection One to two microliters of each PCR product was mixed with 0.5 ml GeneScan-400HD (ROX) internal lane standard (Applied Biosystems, Darmstadt, Germany) and 14.5 ml of deionized formamide (Sigma, Taufkirchen, Germany). The mixture was subjected to heat denaturation in the PCR thermocycler for 3 min at 96 8C. After cooling on ice the samples were injected electrocinetically for 5 s. Detection was performed on a 310 ABI Prism Genetic analyzer according to the manufacturers recommendations (Applied Biosystems, Darmstadt, Germany). Fragment sizes and amount of PCR products were determined automatically applying GeneScan Analysis Software 3.1 (Applied Biosystems, Darmstadt, Germany). 2.4. Validation procedures

2.2. PCR amplification Pentaplex PCR was carried out in a 10 ml reaction mix containing 20 mM Tris–HCl (pH 8.4), 50 mM KCl, 200 mM each dNTP (dATP, dGTP, dCTP, dUTP), 2.2 mM MgCl2, 500 mg/ml bovine serum albumin, 1% Tween 20, 200 nM each of amelogenin primer, 600 nM each of D8S1179 primer, 1000 nM each of vWA primer, 75 nM each of TH01 primer, 150 nM each of D3S1358 Primer, 1 U Platinum Taq DNA Polymerase (Invitrogen, Karlsruhe, Germany), 0.1 U UNG (MBI Fermentas, Leon-Rot, Germany) and a variable amounts of template DNA. To avoid contaminating PCR products samples were digested with UNG (Uracil-DNA Glycosylase) prior to amplification by

2.4.1. Primer specificity Buccal swabs were collected from 200 unrelated healthy volunteers and DNA was immediately extracted using the Chelex 100 method [23]. For comparison 200 human DNA profiles obtained by the ShoP-PCR were compared with profiles applying the SGM und PowerPlex 16 system of the same individuals. To test human specificity DNA from various animals (cat, dog, pig, horse, cow, mouse, rat, frog, fish, sheep, rabbit, guinea pig, goat, deer, black deer, fox, pigeon and herring) and the three closest related primates (gorilla, orangutan, chimpanzee) was isolated. DNA extraction was performed using the QIAamp

Table 1 Description, repeat number of allelic ladder components, allelic size range and primer sequences of the ShoP-PCR loci System

GenBank1 accession

Repeat numbers of allelic ladder components

Size range of allelic ladder components

Primers (50 ! 30 )

Labelled with

Amelogenin

3 bp difference

X: 79 bp, Y: 82 bp

7–18

83–127 bp

VWA

M25858

10–22

93–141 bp

TH01

D00269

4–9, 9.3, 10–11, 13.3

68–107 bp

D3S1358

11449919

12–20

113–145 bp

CCTTTGAAGTGGTACCAGAGCA (forward), GCATGCCTAATATTTTCAGGGAA (reverse) TTTTTGTATTTCATGTGTACATTCG (forward), TCCTGTAGATTATTTTCACTGTGG (reverse) GAATAATCAGTATGTGACTTGGATTGA (forward), GATGATAAATACATAGGATGGATGGA (reverse) GCCTGTTCCTCCCTTATTTCC (forward), AGGTCACAGGGAACACAGACTC (reverse) ACTGCAGTCCAATCTGGGTGAC (forward), GAAATCAACAGAGGCTTGCATG (reverse)

HEX

D8S1179

M55418 and M55419 GO8710

Sequences of the D8S1179 forward primer have been already published [38].

6-FAM HEX NED NED

C. Meissner et al. / Forensic Science International 166 (2007) 121–127

Mini DNA kit (Qiagen, Hilden, Germany) and stored at 20 8C. Additionally, a panel of different bacterial or fungal DNA extracts were applied to proof human-specific annealing of primers. One nanogram DNA of each extract was subjected to PCR as described above. 2.4.2. Sensitivity and DNA mixtures Ten different dilutions of DNA extracts of blood were prepared with an amount of template of 100, 50, 25, 12.5 and 6.25 pg, respectively. The amount of template subjected to the mixture was determined applying the PicoGreen dsDNA Quantification Kit (MoBiTec, Goettingen, Germany) in a VersaFluorTM Fluorometer System (Bio-Rad, Munich, Germany) down to 0.025 ng and then diluted to 0.0125 and 0.00625 ng. Mixture samples (1:0, 19:1, 9:1, 4:1, 2:1, 1:1, 1:2, 1:4, 1.9, 1:19 and 0:1) were prepared of two different DNA extracts with a total input of 1 ng DNA and quantification was performed as described. 2.4.3. DNase I test To monitor successful DNA typing of degraded DNA a DNase I test was performed. High molecular weight DNA was extracted using the MasterPureTM DNA Purification Kit (Epicentre, Madison, WI, USA). Eighty micrograms of the extract were digested with 8 U DNase I in a 200 ml reaction mix applying the DNA-free kit (Ambion, Huntingdon, UK). An aliquot of 20 ml was taken after 0.5, 1, 2, 3, 4, 8, 15 and 30 min, respectively and reaction stopped by adding 4 ml of DNase Inactivation Reagent. 0.2 ml of degraded DNA was taken and subjected to a 10 ml PCR as described above. 2.5. Forensic casework At last our pentaplex approach was tested for suitability in forensic casework analysis. As an example eight poor quality DNA samples were investigated in our study. The samples

123

include a stamp on a letter in a case of blackmail, a diaper, a slip wrapped in a plastic bag and a carpet in a case of sexual abuse, skeletal remains buried in the soil and paraffin-embedded cerebellar tissue of two murdered victims, a single telogen shed hair in a case of an assault of a supermarket and a cigarette butt from a bus stop in an identification case. 3. Results When designing a new multiplex system it is of utmost importance to establish a robust and reliable DNA typing system for all kinds of samples. The ShoP-PCR method was designed to make even amplification of highly degraded DNA possible. This multiplex approach includes the loci amelogenin, TH01, VWA, D3S1358, and D8S1179. None of the alleles described in the literature till date exceeds an allele length of more than 150 bp. Allele and size ranges of each locus are listed in Table 1. The amelogenin primers target a different region than the primer pair commonly used by the forensic community [24]. Our amelogenin primers are 1077 bp away from the region used in commercial kits and were originally described by [25]. As indicated in Table 1 this new target region exhibits a 3 bp difference between the X and Y alleles. PCR conditions and primer concentrations of our ShoP-PCR reaction mix were carefully adjusted to produce optimal results. Adjustment of ramping speeds was an important step, not only for the GeneAmp 2400 thermal cycler, but for other cyclers (GeneAmp 9600, Applied Biosystems, Darmstadt, Germany; PCR Express, Thermo Electron/Hybaid, Dreieich, Germany) either. To exclude mutations of primer binding sequences 200 healthy unrelated individuals were tested using our ShoP-PCR and the SGMplusTM Kit (Applied Biosystems, Darmstadt, Germany) or PowerPlex 16 SystemTM (Promega, Mannheim, Germany). Except in one case where allele 16 was not detectable in a 16/17 genotype of the VWA locus applying the SGMplusTM Kit, no differences were observed between the typing results of the

Table 2 Peak height of stutters in percentage (%) with S.E. (standard error) Allele (n)

6 7 8 9 9.3 10 11 12 13 14 15 16 17 18 19 20

TH01 (n = 44)

D8S1179 (n = 107)

D3S1358 (n = 123)

Stutter (%)

S.E. (%)

Stutter (%)

S.E. (%)

3.64 2.37 4.82 3.50 4.03 12.99

2.20 0.62 2.54 0.97 2.79

3.65 (4)

1.77

5.89 6.22 6.38 7.08 6.58 8.02 9.48

1.31 2.12 2.03 2.39 2.41 2.09 3.19

(10) (4) (8) (11) (10) (1)

(16) (8) (16) (33) (14) (13) (3)

VWA (n = 102)

Stutter (%)

S.E. (%)

Stutter (%)

S.E. (%)

7.69 8.89 9.56 10.30 11.16 9.77

2.03 2.42 2.60 2.71 2.53 0.44

2.46 6.08 6.53 7.15 8.67 10.46 7.40

1.25 2.73 1.80 2.05 2.90 2.00 0.14

(17) (35) (24) (24) (20) (3)

(11) (16) (22) (25) (17) (9) (2)

n is the number of observed and detectable allelic stutters; values in parentheses represent number of observations for each stutter product.

124

C. Meissner et al. / Forensic Science International 166 (2007) 121–127

Fig. 1. DNA typing results of the ShoP-PCR applying 100 pg (a), 25 pg (b), 12.5 pg (c) and 6.25 pg (d) of template DNA. Note allelic drop out in the panel at the bottom (d) for the loci amelogenin, TH01 and VWA, respectively.

commercially available kits or the ShoP-PCR. Frequencies of the 200 unrelated individuals were as expected according to the Hardy–Weinberg equilibrium. When applying 200 pg of DNA template, the four autosomal loci displayed differences of heterozygote imbalance. The lowest value was observed for the TH01 system with 4.07  11.30%, the highest value for D8S1179 with 13.38  13.29%. D3S1358 displays a heterozygote imbalance of 9.35  9.40%, VWA 7.23  10.64%.

The occurrence of allele-specific stutter bands is described in Table 2. Whereas peak height of stutters was similar for each allele independent of its size for the TH01 locus, peak height increases in dependence of size of the amplicons for the VWA and especially the D3S1358 loci, while D8S1179 displays no tendency. When a panel of different animal, bacterial and fungal DNA samples was applied, no signals were detected, except

Fig. 2. DNA as artificially degraded by digestion with DNase I for 2 min (a and d), 3 min (b and e) and 4 min (c and f). DNA profiles using the SGMplusTM Kit (A) and the ShoP-PCR (B). A complete profile of the five ShoP loci was obtained after 3 min and four of five loci were clearly visible after 4 min.

C. Meissner et al. / Forensic Science International 166 (2007) 121–127

125

Fig. 3. DNA tying result of the ShoP-PCR of DNA extracted from a slip (A) and a diaper (B) in a case of suspected sexual child abuse. Commercially available kits failed to display any signals.

Table 3 DNA typing results of eight different forensic materials Sample

Material

Age

ShoP

Commercial kits

1 2 3 4 5 6 7 8

Telogen shed hair Bone Slip Pampers Carpet Stamp Cigarette butt Paraffin-embedded tissue

2 years 7 years 2 years 7 days Unknown 6 months Unknown 15 years

5/5 3/5 5/5 5/5 5/5 4/5 5/5 4/5

0/5 0/5 1/5 0/5 0/5 0/5 2/5 2/5

orangutan and gorilla revealed some signals. Chimpanzee DNA displayed a profile with amelogenin, TH01, D3S1358 and VWA primers, albeit the specific signals for VWA were out of the range of the human ladder (data not shown). Different concentrations of template were tested and revealed a detection limit of 12.5 pg template DNA extracted from postmortem tissue without signs of decomposition. This equals the amount of roughly two diploid cells. When 6.25 pg were applied, allelic drop out occurs as expected (Fig. 1). Mixtures of 50 pg in a total of 1000 pg DNA (1:19) were clearly detectable (data not shown). When a DNase I test was performed, a complete DNA profile was obtained after a digestion of 3 min, whereas the SGMplusTM Kit displays only signals for the loci amelogenin, D3S1358 and D8S1179. At 4 min digestion four of five ShoP loci were clearly visible in contrast to the SGMplusTM Kit, where only locus D19S433 was detectable (Fig. 2A and B). As indicated in Table 3 our ShoP-PCR analysis revealed in forensic casework analysis for all eight samples a better result in comparison to the commercial kits. Examples of ShoP-PCR profiles are presented for a slip (Fig. 3A) and a diaper (Fig. 3B) in a case of sexual abuse. In five of the cases a complete DNA profile was detectable for the ShoP loci, in cases 6 and 8 four of the five loci were clearly visible. Even the sample of bone buried in the soil for 7 years displayed three complete loci. In comparison the SGMplusTM and/or the PowerPlex 16 SystemTM revealed less or even no results. 4. Discussion The short pentaplex PCR includes the loci amelogenin, TH01, VWA, D3S1358 and D8S1179. It was designed to allow successful DNA typing even in cases with highly degraded

DNA or a low amount of template. Because even poor quality DNA reveals fragment sizes of 100–150 bp [11,14,15], none of the common alleles of the five loci exceeds a fragment length of more than 150 bp. To avoid an overlap of alleles of two different loci labelled by the same fluorescent dye, primers were chosen with a 10 bp gap between the smallest common allele of one locus and largest of the other. Some samples revealed interfering peaks, which exhibit a broader appearance, easy to distinguish from specific dye-labelled PCR products. In other cases a few artificial peaks or dye blobs were detected, but artefact activity did not lead to difficulties in interpretation of results, because most of them were outside the fragment length of the true alleles of a locus or beneath the detection limit of 50 RFU. When designing primers for PCR amplification, sequence alignment procedures must exclude primer binding sites other than that of interest. Polymorphisms within primer binding regions may lead to the presence of a null allele [26]. This phenomenon could result in the occurrence of false homozygotes, as has been recently described for the D5S818 locus, where allele 10 segregates with a mutation of the primer binding site [27,28]. In PCRs with short target amplicons the closest position of primers to the STR of interest is not always that of choice, because flanking regions of primers often contain repeat structures or sequences, which could hamper the PCR by avoiding primer annealing or generating PCR artefacts [20]. So each of the designed PCRs should be carefully tested for artefacts and reliability in a large panel of human samples, animal DNA as well as forensic samples. Allelic drop out of the five loci was tested comparing the DNA profile of 200 individuals with results of the two commercially available kits PowerPlex 16 SystemTM (Promega, Madison, USA) and SGMplusTM Kit (Applied Biosystems, Darmstadt, Germany). In all cases identical profiles were obtained and no preferential amplification, allelic or loci drop out has been detected. The loss of allele 16 observed in the locus VWA applying the SGMplusTM Kit is supposed to be a result of a mutation in the primer binding site as had been already observed for allele 19 [29], because PowerPlex 16 SystemTM and ShoP-PCR profiles were identical. Specificity of the primers was tested applying a panel of animal, bacterial and fungal DNA. Only primate DNA revealed some detectable signals, partly off-ladder alleles, which has been already observed in other studies of human STR-loci, when primate DNA was tested [30,31].

126

C. Meissner et al. / Forensic Science International 166 (2007) 121–127

In dilution experiments at least 12.5 pg of human DNA produced reliable results, albeit heterozygosity imbalance was observed in some cases. It is well known that under conditions when only a small amount of DNA is available, a tendency for preferential amplifications occurs [32] and allelic imbalance and allelic drop out should be kept in mind. The short fragments of the ShoP-PCR increase sensitivity due to a higher amplification efficiency and a complete profile was detectable even when amounts of 12.5 pg were applied (Fig. 1). When 6.25 pg template DNA were subjected to PCR, allelic drop out was observed, because stochastic effects of primer annealing and PCR amplification are expected. In other multiplex PCRs these phenomena are observed when applying DNA amounts of 200 pg [33]. Another important issue to address in validation procedures is the reliability and sensitivity when mixed stains are analysed, because the limit of detection of the minor component in mixed DNA extracts depends on the amount of this component. Using a total of 1000 pg template, it was possible to obtain a complete profile of an amount of 50 pg DNA. The detectable limit of 1:19 of the minor component is similar to other multiplex approaches [33,34]. The major advantage of our system becomes quite clear, when degraded material has been analysed. Degraded DNA was prepared by DNase I digestion under controlled conditions. Even after digestion of 4 min four of five loci were clearly visible. It is not surprising that allelic drop out of the largest locus, D3S1358 occurs at first. Failure of signal detection for the locus FGA was observed in a similar manner when the AmpFISTRTM Blue PCR Amplification Kit was applied and two of three loci were detectable [34]. Our approach allows typing of degraded DNA due to the small size of the target fragments, because none of the detectable alleles exceeds 150 bp in length. A disadvantage of our ShoP-PCR is the relatively low power of discrimination with 1:3.16  104, because only four autosomal loci are displayed. In comparison, the Power of discrimination for the SGMplusTM Kit (1:3.3  1012) and the PowerPlex 16 System (1:1.83  1017) are orders of magnitude higher. So performance of commercial kits is better in cases when amount of intact DNA is sufficient. Because the aim of our study was to present a PCR useful for degraded material in forensic casework analysis, the ShoP-PCR was tested on a panel of forensic samples like telogen shed hair, bone buried in the soil or formalin-embedded tissues, which revealed only up to two loci, when SGMplusTM Kit or PowerPlex 16 SystemTM were applied. Telogen shed hairs are often found at crime scenes. Due to its nature, DNA enclosed in the keratin matrix is intensively degraded and typing of shed hairs without a bulb often reveals no results [5]. As indicated in Table 3 a single telogen shed hair collected from the clothes of a victim displayed the complete profile. DNA extracted from human bones buried in the soil or found in damp environment is often extensively degraded [13,16]. When typing DNA extracts of bones of air crash victims it was observed, that allelic or loci drop out occurs, when alleles were longer than 200 bp [35]. As has been shown for bone buried in the soil (case 2), our ShoPPCR approach will lead to successful results, due to the small

size of target fragments. Formalin-fixation induces degradation of DNA, even after fixation times of 3 h [36,37]. In routine pathology tissue is often fixed for at least 2 days in buffered or unbuffered formalin before paraffin embedding. Applying formalin-fixed, paraffin-embedded tissue (case 8) after a storage time of 15 years a complete DNA profile was obtained, so this PCR will be useful when only paraffin-embedded material is available, which is often stored over decades of time. 5. Conclusion In cases with an amount of DNA of more than 12.5 pg or in cases when only highly degraded DNA is available our system allows successful DNA typing. The ShoP-PCR approach seems to be extremely useful in typing of telogen hairs, highly putrefied tissue or tissue fixed in formalin for longer periods of time. The design of new mini primer pairs and their combination in multiplex PCRs provide improved results when allelic drop out occurs due to degraded DNA or due to mutations of primer binding sites. In future, it will be important to have more than a handful of different multiplex mini-STRs to ensure optimal results in forensic casework analysis. References [1] P. Hoff-Olsen, S. Jacobsen, B. Mevag, B. Olaisen, Microsatellite stability in human post-mortem tissues, Forensic Sci. Int. 119 (2001) 273–278. [2] C. Kimpton, D. Fisher, S. Watson, M. Adams, A. Urquhart, J. Lygo, P. Gill, Evaluation of an automated DNA profiling system employing multiplex amplification of four tetrameric STR loci, Int. J. Legal Med. 106 (1994) 302–311. [3] R. Schoske, P.M. Vallone, C.M. Ruitberg, J.M. Butler, Multiplex PCR design strategy used for the simultaneous amplification of 10 Y chromosome short tandem repeat (STR) loci, Anal. Bioanal. Chem. 375 (2003) 333–343. [4] J.M. Butler, J.M. Devaney, M.A. Marino, P.M. Vallone, Quality control of PCR primers used in multiplex STR amplification reactions, Forensic Sci. Int. 119 (2001) 87–96. [5] A. Hellmann, U. Rohleder, H. Schmitter, M. Wittig, STR typing of human telogen hairs—a new approach, Int. J. Legal Med. 114 (2001) 269–273. [6] A. Alvarez Garcia, I. Munoz, C. Pestoni, M.V. Lareu, M.S. RodriguezCalvo, A. Carracedo, Effect of environmental factors on PCR-DNA analysis from dental pulp, Int. J. Legal Med. 109 (1996) 125–129. [7] J. Burger, S. Hummel, B. Hermann, W. Henke, DNA preservation: a microsatellite-DNA study on ancient skeletal remains, Electrophoresis 20 (1999) 1722–1728. [8] M.M. Holland, D.L. Fisher, L.G. Mitchell, W.C. Rodriquez, J.J. Canik, C.R. Merril, V.W. Weedn, Mitochondrial DNA sequence analysis of human skeletal remains: identification of remains from the Vietnam War, J. Forensic Sci. 38 (1993) 542–553. [9] P. Hoff-Olsen, B. Mevag, E. Staalstrom, B. Hovde, T. Egeland, B. Olaisen, Extraction of DNA from decomposed human tissue. An evaluation of five extraction methods for short tandem repeat typing, Forensic Sci. Int. 105 (1999) 171–183. [10] S. Banaschak, A. Du Chesne, B. Brinkmann, Multiple interchanging of tissue samples in cases of breast cancer, Forensic Sci. Int. 113 (2000) 3–7. [11] M. Ho¨ss, P. Jaruga, T.H. Zastawny, M. Dizdaroglu, S. Paabo, DNA damage and DNA sequence retrieval from ancient tissues, Nucl. Acids Res. (7) (1996) 1304–1307. [12] K. Bender, M.J. Farfan, P.M. Schneider, Preparation of degraded human DNA under controlled conditions, Forensic Sci. Int. 139 (2004) 135–140.

C. Meissner et al. / Forensic Science International 166 (2007) 121–127 [13] M. Graw, H.J. Weisser, S. Lutz, DNA typing of human remains found in damp environments, Forensic Sci. Int. 113 (2000) 91–95. [14] S. Pa¨a¨bo, Ancient DNA: extraction, characterization, molecular cloning, and enzymatic amplification, Proc. Natl. Acad. Sci. U.S.A. 86 (1989) 1939–1943. [15] O. Handt, M. Krings, R.H. Ward, S. Pa¨a¨bo, The retrieval of ancient human DNA sequences, Am. J. Hum. Genet. 59 (1996) 368–376. [16] D.L. Fisher, M.M. Holland, L. Mitchell, P.S. Sledzik, A.W. Wilcox, M. Wadhams, V.W. Weedn, Extraction, evaluation, and amplification of DNA from decalcified and undecalcified United States Civil War bone, J. Forensic Sci. 38 (1993) 60–68. [17] C. Lassen, S. Hummel, B. Herrmann, Comparison of DNA extraction and amplification from ancient human bone and mummified soft tissue, Int. J. Legal Med. 107 (1994) 152–155. [18] G.G. Shutler, P. Gagnon, G. Verret, H. Kalyn, S. Korkosh, E. Johnston, J. Halverson, Removal of a PCR inhibitor and resolution of DNA STR types in mixed human-canine stains from a five year old case, J. Forensic Sci. 44 (1999) 623–626. [19] M. Strom, S. Rechitsky, Use of nested PCR to identify charred human remains and minute amounts of blood, J. Forensic Sci. 43 (1998) 696–700. [20] J.M. Butler, Y. Shen, B.R. McCord, The development of reduced size STR amplicons as tools for analysis of degraded DNA, J. Forensic Sci. 48 (2003) 1054–1064. [21] P. Wiegand, M. Kleiber, Less is more—length reduction of STR amplicons using redesigned primers, Int. J. Legal Med. 114 (2001) 285–287. [22] S. Rozen, H. Skaletsky, Primer3 on the WWW for general users and for biologist programmers, Meth. Mol. Biol. 132 (2000) 365–386. [23] P.S. Walsh, D.A. Metzger, R. Higuchi, Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material, Biotechniques 10 (1991) 506–513. [24] K.M. Sullivan, A. Mannucci, C.P. Kimpton, P. Gill, A rapid and quantitative DNA sex test: fluorescence-based PCR analysis of X–Y homologous gene amelogenin, Biotechniques 15 (1993) 636–638, 640–641. [25] H. Haas-Rochholz, G. Weiler, Additional primer sets for an amelogenin gene PCR-based DNA-sex test, Int. J. Legal Med. 110 (1997) 312–315.

127

[26] K. Lazaruk, J. Wallin, C. Holt, T. Nguyen, P.S. Walsh, Sequence variation in humans and other primates at six short tandem repeat loci used in forensic identity testing, Forensic Sci. Int. 119 (2001) 1–10. [27] P. Kuzniar, R. Ploski, STR data for the power plex-16 loci in a population from Central Poland, Forensic Sci. Int. 139 (2004) 261–263. [28] M. Delamoye, C. Duverneuil, K. Riva, M. Leterreux, S. Taieb, P. De Mazancourt, False homozygosities at various loci revealed by discrepancies between commercial kits: implications for genetic databases, Forensic Sci. Int. 143 (2004) 47–52. [29] M.C. Kline, B. Jenkins, S. Rodgers, Non-amplification of a VWA allele, J. Forensic Sci. 43 (1998) 250. [30] C.A. Crouse, J. Schumm, Investigation of species specificity using nine PCR-based human STR systems, J. Forensic Sci. 40 (1995) 952–956. [31] E. Meyer, P. Wiegand, S.P. Rand, D. Kuhlmann, M. Brack, B. Brinkmann, Microsatellite polymorphisms reveal phylogenetic relationships in primates, J. Mol. Evol. 41 (1995) 10–14. [32] D. Kloosterman, P. Kersbergen, Efficacy and limits of genotyping low copy number (LCN) DNA samples by multiplex PCR of STR loci, J. Soc. Biol. 197 (2003) 351–359. [33] A. Junge, T. Lederer, G. Braunschweiger, B. Madea, Validation of the multiplex kit genRESMPX-2 for forensic casework analysis, Int. J. Legal Med. 117 (2003) 317–325. [34] J.M. Wallin, M.R. Buoncristiani, K.D. Lazaruk, N. Fildes, C.L. Holt, P.S. Walsh, TWGDAM validation of the AmpFISTR blue PCR amplification kit for forensic casework analysis, J. Forensic Sci. 43 (1998) 854–870. [35] W. Goodwin, A. Linacre, P. Vanezis, The use of mitochondrial DNA and short tandem repeat typing in the identification of air crash victims, Electrophoresis 20 (1999) 1707–1711. [36] Y. Tokuda, T. Nakamura, K. Satonaka, S. Maeda, K. Doi, S. Baba, T. Sugiyama, Fundamental study on the mechanism of DNA degradation in tissues fixed in formaldehyde, J. Clin. Pathol. 43 (1990) 748–751. [37] N. Yagi, K. Satonaka, M. Horio, H. Shimogaki, Y. Tokuda, S. Maeda, The role of DNase and EDTA on DNA degradation in formaldehyde fixed tissues, Biotech. Histochem. 71 (1996) 123–129. [38] M.D. Barber, B.H. Parkin, Sequence analysis and allelic designation of the two short tandem repeat loci D18S51 and D8S1179, Int. J. Legal Med. 109 (1996) 62–65.

Forensic Science International 166 (2007) 128–138 www.elsevier.com/locate/forsciint

LoComatioN: A software tool for the analysis of low copy number DNA profiles Peter Gill a,*, Amanda Kirkham a, James Curran b b

a Forensic Science Service, Trident Court, 2960 Solihull Parkway, Solihull B37 7YN, UK Department of Statistics, The University of Auckland, Private Bag 92019, Auckland, New Zealand

Received 23 February 2006; accepted 11 April 2006 Available online 8 June 2006

Abstract Previously, the interpretation of low copy number (LCN) STR profiles has been carried out using the biological or ‘consensus’ method— essentially, alleles are not reported, unless duplicated in separate PCR analyses [P. Gill, J. Whitaker, C. Flaxman, N. Brown, J. Buckleton, An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA, Forens. Sci. Int. 112 (2000) 17–40]. The method is now widely used throughout Europe. Although a probabilistic theory was simultaneously introduced, its time-consuming complexity meant that it could not be easily applied in practice. The ‘consensus’ method is not as efficient as the probabilistic approach, as the former wastes information in DNA profiles. However, the theory was subsequently extended to allow for DNA mixtures and population substructure in a programmed solution by Curran et al. [J.M. Curran, P. Gill, M.R. Bill, Interpretation of repeat measurement DNA evidence allowing for multiple contributors and population substructure, Forens. Sci. Int. 148 (2005) 47–53]. In this paper, we describe an expert interpretation system (LoComatioN) which removes this computational burden, and enables application of the full probabilistic method. This is the first expert system that can be used to rapidly evaluate numerous alternative explanations in a likelihood ratio approach, greatly facilitating court evaluation of the evidence. This would not be possible with manual calculation. Finally, the Gill et al. and Curran et al. papers both rely on the ability of the user to specify two quantities: the probability of allelic drop-out, and the probability of allelic contamination (‘‘drop-in’’). In this paper, we offer some guidelines on how these quantities may be specified. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Low copy number (LCN); Automation; LoComatioN; Likelihood ratio; Propositions

1. Introduction Low copy number (LCN) DNA profiling is a term used to describe the analysis of very small amounts of DNA from a few cells (<200 pg). In ideal conditions it is possible to successfully get a profile from a single cell (6 pg) by raising the number of PCR amplification cycles from 28 to 34 [1]. Although the sensitivity of the test is greatly improved, Gill et al. [2] and Whitaker et al. [3] showed that interpretation is complicated. One phenomenon is that heterozygotes become highly ‘‘imbalanced’’. Heterozygote imbalance arises when one of the alleles in a heterozygous genotype amplifies more strongly than the other, even though one person contributed both of the alleles. We observe this ‘‘imbalance’’ as a difference in peak * Corresponding author. E-mail address: [email protected] (P. Gill). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.04.016

heights or peak areas. Secondly, an extreme form of heterozygote imbalance results in disappearance of an allele, giving the false appearance of a homozygote. This phenomenon is known as allele ‘drop-out’. Finally, ‘drop-in’ (contamination), where one or two additional alleles can appear in the profile and must be included in the assessment [4]. LCN profiles are often DNA mixtures. In modern DNA mixture interpretation it has become customary to use peak height or area information to guide the choice of feasible genotype combinations. However, interpretation methods that include a consideration of peak height/area are not appropriate for LCN evidence. These ‘‘quantitative’’ methods have primarily been developed for profiles where there is a significant amount (>200 pg) of DNA present. The assumption that the peak height/area of alleles is proportional to the actual amount of DNA present [5–7] is well established, however with LCN, stochastic effects compromise this [8].

P. Gill et al. / Forensic Science International 166 (2007) 128–138

Although a probabilistic method has been published [9,10] the likelihood ratio (LR) calculations are far too complex to carry out manually, especially when the theory was extended to include mixture interpretation with multiple contributors. An interim method, called the ‘‘biological model’’, was introduced. The biological model depended upon the derivation of a ‘‘consensus’’ profile. A consensus profile only reports alleles that were reproducible from two or more replicate analyses of an extracted DNA sample [9,11]. As contamination tended to be a single tube event of low probability, it was unlikely that these alleles would be replicated in different analyses and reported in the consensus profile. The biological model tended to behave in a conservative way relative to the formal statistical model, but does not make full use of the information available in the replicate DNA profiles. Curran et al. [12] recently introduced a set theoretic formalization to allow the automatic calculation of LRs for LCN profiles. This method has been implemented in a fully functional software application called LoComatioN. LoComatioN is a hypothesis driven expert system that enables LRs for any number of different LCN propositions to be evaluated. The construction of the LR follows the standard format, requiring an evaluation of the probability of observing the evidence under the prosecution and the defence hypotheses, Hp and Hd, respectively. We call these hypotheses ‘‘propositions’’. An example might be a rape case where a woman alleges she was raped by exactly one man. The prosecution proposition (Hp) is that the crime scene stain consists of the victim (V) and the suspect (S). The alternative or defence proposition (Hd) is that the victim and someone unrelated to the suspect were the only contributors. We denote this V + unknown (U). Of course, more complex propositions may be suggested by the defence, and it may be desirable to evaluate the LR with respect to several different pairs of propositions. Although the theory to analyse different propositions exists, in practice the computational requirements for a reporting officer doing the calculations manually are very time-consuming (and therefore potentially error prone). As a result this option is often precluded. This inability to provide adequate calculations to the court for multiple propositions is a limiting factor and might be detrimental because cases may be reported as ‘‘inconclusive’’. The advantage of LoComatioN, is that the scientist is able to input data from up to five replicate analyses, and is able to consider up to five contributors to any mixture where the propositions can be altered at will. This means that for virtually all mixtures, the scientist can now rapidly evaluate any number of propositions that the court requires. We hope that this means the ‘‘inconclusive’’ category will become something of the past. The ability to evaluate multiple propositions means that LoComatioN has an important role as an exploratory tool. We show how sensitive the LR is to different conditioning statements/propositions by reference to a complex case. To facilitate the court going process and to resolve potential uncertainties about the effects of different conditioning statements, we have introduced guidance to formulate propositions by incorporating some generalisations of Brenner et al. [13], Weir [14] and Buckleton et al. [15].

129

2. Formulation of propositions We use the following notation to show the respective propositions in a typical mixture case conditioned on a victim (V), suspect (S) and unknown (U) where the propositions are Hp: V + S and Hd: V + U: LR ¼

PrðEjHp Þ PrðEjHd Þ

where the likelihood ratio is comprised of Hp (the prosecution proposition) in the numerator and Hd (the defence proposition) in the denominator. E is the evidence of the DNA crime profile. The prosecution proposition (Hp) is initially based upon the testimony of witnesses and other circumstances of the case. DNA profiling is carried out on a crime stain and the results are used to confirm or to refute the proposition. If the profile matches the suspect (S), then the proposition Hp is supported. In a DNA mixture, alleles that match S may be present, providing support for Hp. However, additional alleles from other sources may also be present and these may provide support for the alternative defence proposition (Hd). Further refinement of propositions might be required [16,17]. It is not always easy to specify propositions in complex cases where multiple perpetrators/victims may be present. The DNA result itself may indicate that different explanations are possible. Furthermore, it is possible that Hp and Hd could be very different from each other. For example under Hp we might consider a victim and suspect to be the contributors (V + S), whereas under Hd we might examine more complex propositions such as three unknowns being the contributors to the stain (U1 + U2 + U3). There is a common misconception that the number of contributors (nc) under Hp and Hd should be the same. They do not. 3. Allele drop-out and the Q designation Drop-out is an important defining feature of LCN. There are two aspects to be included in probabilistic calculations: the first is to estimate the probability of drop-out Pr(D) and the second is to include the dropped out allele in the probabilistic assessment. Originally the F designation [9] was used to signify the possibility of drop-out event; a sample that shows a single allele, a, can be designated aF, where F can be any allele, including a. The probability of F = 1 since it includes all allelic possibilities, the probability of a is pa, hence Pr(aF) = 2pa. However, this formula may not be conservative, i.e. can over estimate the LR in favour of Hp [9]. This is more likely to happen when the probability of drop-out is low. We have introduced an improved concept into LoComatioN to facilitate programming. If drop-out is required to support a proposition we consider that the identity of the unknown allele Q can be anything, except those already observed in the DNA profile: PrðQÞ ¼ 1 

n X i¼1

pi

130

P. Gill et al. / Forensic Science International 166 (2007) 128–138

where n alleles are observed in the profile and pi is the frequency of the ith allele. Consider the following simple example. The crime stain profile at locus THO1 has one allele of type 11. The suspect (S) is genotype 9,11. Under Hp, we argue that allele 9 must have dropped out. Under Hd, evaluation of the alternative explanation (U) would include a probabilistic determination of all possible pairwise combinations (that must include allele 11): 4,11; 5,11; 6,11, etc. a total of nine different combinations to be computed. The Q designation is used, given drop-out, where Pr(Q)=pQ = 1  p11, and this achieves exactly the same result in just one computational step. The combination p211 is included under the hypothesis that no drop-out has occurred. When mixtures are considered, the computational savings are even greater. 3.1. Using the Q designation to formulate Hp and Hd As an example, if the stain profile E = abc; S = ab; nc = 2 and all three alleles are low level, then under Hp, if drop-out has occurred we consider all pairwise combinations of cQ where p Q = 1  p a  p b  p c: PrðEjHp ; DÞ ¼ PrðEjHp ; DÞPrðDÞ;

Alternatively, if no drop-out has occurred: ¯ ¼ PrðEjHp ÞPrðDÞ; ¯ PrðEjHp ; DÞ PrðEjHp Þ ¼ p2c þ 2 pa pc þ 2 pb pc

(2)

Hence, Pr(EjHp) comprises the sum of terms (1) and (2). Under Hd, with two unknown (U1, U2) contributors, given drop-out: PrðEjHd Þ ¼ 24 pa pb pc pQ (3)

With no drop-out, such that alleles a, b, c are shared between two contributors: ¯ ¼ PrðEjHd ÞPrðDÞ; ¯ PrðEjHd ; DÞ PrðEjHd Þ ¼ 12 pa pb pc ð pa þ pb þ pc Þ

(4)

The likelihood ratio is LR = Pr(EjHp)/Pr(EjHd): LR ¼

A contaminant event is the spurious occurrence of single alleles from multiple sources, assumed to be independent events. Probability of contamination is estimated from negative controls as described by Gill and Kirkham [4]. Laboratory records d ¼ 0:05 per sample where indicate a level of approximately PrðCÞ d ¼ n PrðCÞ LN where n is the number of alleles observed in a series of negative controls and N the total number of negative controls analysed and L is the number of loci tested per sample (whether or not alleles are actually observed). The ‘‘hat’’ over Pr(C) indicates that this is an estimate. The probability of any given allele appearing as a contaminant is approximated to be the same as the probability of its occurrence in the white Caucasian population (from a frequency database). 6. A fully worked example with drop-out and contamination

PrðEjHp Þ ¼ 2 pc pQ (1)

PrðEjHd ; DÞ ¼ PrðEjHd ÞPrðDÞ;

5. Estimation of Pr(C)

¯ PrðDÞð2 pa þ 2 pb þ pc Þ þ PrðDÞð2 pQ Þ ¯ pa þ pb þ pc Þ þ PrðDÞð2 pQ Þ 12 pa pb ½PrðDÞð

(5)

A suspect’s genotype at a particular locus is ab. The crime sample profile (E) is a. The prosecution proposition (Hp) states that the suspect (S) is the offender. This can only be explained if drop-out of allele b had occurred. The defence proposition (Hd) is that the offence has been committed by an unknown individual (U), unrelated to the suspect. Using our previously defined notation the likelihood ratio using propositions Hp: S and Hd: U is LR ¼

PrðEjHp Þ PrðEjHd Þ

Formulae for the numerator and denominator are given in Table 1, illustrating use of the Q virtual allele designation in conjunction with the probability of drop-out, Pr(D) and the ¯ ¼ 1  PrðCÞ. probability of no contamination PrðCÞ The calculations for this simple example are just about manageable by hand, but most propositions will be much more complicated than this, comprising mixtures from two or more people and two or more replicates. An example of LoComatioN output and associated statistical analysis is given in Appendix II. 7. Casework example to illustrate evaluation of multiple propositions

4. Estimation of Pr(D)

7.1. Case circumstances

From Gill et al. [8], for low copy number DNA, in the absence of degradation, it is reasonable to assume that the chance of allele drop-out is independent of the locus. Note that if significant degradation has occurred then high molecular weight loci will be affected preferentially. Under LCN conditions, where DNA is amplified 34 cycles, the biochemistry/detection system will distinguish a single copy of DNA at any SGM+ locus [1]. We provide a method to estimate Pr(D) by simulation based on the assumption that Pr(D) is constant across all loci (Appendix I).

Late one night, two cohabiting females were woken by a masked man who had broken into their flat. The intruder threatened the women with a hammer. He ordered them to engage in sexual acts but the victims did not comply. One shouted for help and the other fought off the assailant. Both victims sustained injuries caused by the hammer. The assailant ran away, discarding the hammer outside the flat, which was subsequently recovered. On questioning, the suspect denied that the hammer was his.

P. Gill et al. / Forensic Science International 166 (2007) 128–138

131

Table 1 An illustration of the correct use of Q when drop-out is considered Suspect (Mj)

Pr(Mj)

Pr(E = ajMj)

Product

Hp numerator calculation a,b

1

¯ ¯ PrðDÞPrðDÞPrð CÞ

¯ ¯ PrðDÞPrðDÞPrð CÞ

Possible random men (Mj)

Pr(Mj)

Pr(E = ajMj)

Producta

Hd denominator calculation a,a a,Q Q,Q

p2a 2papQ p2Q

¯ ¯ 2 PrðCÞ PrðDÞ ¯ ¯ PrðDÞPrðDÞPrðCÞ Pr(D)2Pr(C)pa

¯ p2a ¯ 2 PrðCÞ PrðDÞ ¯ pa pQ ¯ 2PrðDÞPrðDÞPrð CÞ 2 PrðDÞ PrðCÞ pa p2Q

The crime stain is of type a, the suspect is genotype ab and under Hp, we assume that given S, allele b has dropped out with probability Pr(D). Under Hd, given that the suspect is innocent, then drop-out may or may not have happened. We evaluate a set of possible ‘‘random man’’ genotypes worth considering M1, M2, M3. a Denominator = sum of the products.

Table 2 Tabulated PCR amplification results from casework example Allelic results observed at each loci tested Amelo D3 Sample (R1) Sample (R2) Victim 1 Victim 2 Suspect

VWA

D16

D2

D8

D21

D18

11 13 14 20 23 24 25 11 12 13 15 28 31

D19

THO

FGA

12 14 15.2 17.2

6 8 9 9.3 22

XY

14 16 15 16 19

XY

14 16 15 16 17 19 11 13 14 20 24 25

11 12 13 15 28 29 30 31 31.2 13 14 16 17 12 13 14 15.2 17.2 6 8 9 9.3 22 23 25

XX XX XY

16 16 15 16 15 17 16 19 14 16 15 19

11 15 11 13 12 13

13 13 12 13 11 14

20 20 18 25 24 25

29 30 29 30 28 31

17 17 15 17 14 17

12 14 14 14 15.2 17.2

68 67 9 9.3

22 25 20 22 22 23

7.2. Propositions and DNA analysis

7.3. Traditional consensus method (biological model)

The overall purpose of the investigation was to establish whether the hammer was relevant evidence—i.e. was the hammer used/not used in the attack? The specific purpose of the DNA investigation was to establish if there was evidence to support or to refute alternative propositions [16] of the kind:

The consensus approach [9] was dependent upon experimental reproducibility of individual alleles. The method compared two separate PCR amplification results and the calculation of the LR was derived from the consensus of duplicated alleles at each locus in R1 and R2 (Table 3). The consensus approach uses the F designation to signify drop-out. The assumptions in this model were:

 Hp: the DNA from the hammer originated from the suspect and two victims;  Hd: the DNA from the hammer originated from an unknown individual unrelated to the suspect, and two victims. The hammer-head was swabbed and two LCN PCR amplification replicates (R1 and R2) were obtained (Table 2). The results showed that at some loci more than two alleles were present, suggesting a mixture (following guidelines of Clayton et al. [6]). Both PCR amplification and extraction reagent negatives were blank, indicating no obvious source of gross contamination. From laboratory records of negative controls, Pr(C) = 0.05.

1. a three person mixture, nc = 3; 2. both victims were considered to be contributors under both Hp and Hd. We evaluate propositions Hp: V1 + V2 + S and Hd: V1 + V2 + U. The standard approach was used: any alleles that matched either of the victims were subtracted to leave a partial profile (Table 4), interpreted as S under Hp and U under Hd. There were seven alleles shared between both victims and the suspect. The F designation was subsequently assigned to

Table 3 Tabulated consensus PCR amplification results from R1 and R2 in the casework example Allelic results (consensus) observed at each loci tested

Consensus result

Amelo

D3

VWA

D16

D2

D8

D21

XY

14 16

15 16 19

11 13 14

20 24 25

11 12 13 15

28 31

NB. The alleles in bold denote alleles that could be attributed to the victims.

D18

D19

THO

FGA

12 14 15.2 17.2

6 8 9 9.3

22

132

P. Gill et al. / Forensic Science International 166 (2007) 128–138

Table 4 Tabulated ‘foreign’ alleles defining the assailant’s DNA components in the DNA mixture taken from the hammer ‘Foreign’ alleles defining the offender’s DNA profile

Offender

Amelo

D3

VWA

D16

D2

D8

D21

D18

D19

THO

FGA

Y

14 F

FF

11 14

24 F

12 F

28 31

FF

15.2 17.2

9 9.3

FF

any locus where one allele was present [9] to signify the possibility of allele drop-out: LR ¼

PrðEjHp Þ PrðEjHd Þ

¼ 1:55  106 ðwhite Caucasian reference databaseÞ 7.3.1. LoComatioN analysis We now evaluate the effect of comparing different alternative pairs of propositions in the context of the fully probabilistic model that incorporates probabilities of drop-out and contamination into the LR [9,12]. This model is much more powerful than the consensus approach, taking the interpretation process a stage further. A consensus profile is not derived. Consequently, it is possible to calculate the LR relative to a single analysis (R1), although replicate (R1, R2, . . ., Rn) analyses are much to be preferred, because more information is incorporated into the calculation. The Q virtual allele is used when drop-out occurs, instead of F in the ‘consensus’ method. 7.4. Application of the theory to evaluate multiple propositions Casework circumstances are often complex. Multiple pairs of propositions may be possible, but the prime consideration is that the suspect S is always in the numerator under Hp and this is replaced by U in the denominator under Hd. A dialogue may ensue in court where the scientist is requested to evaluate the LR using multiple ‘what-if’ propositions. LoComatioN can be used as an exploratory tool for this purpose. The profile in the example can be interpreted using several different propositions conditioned on nc = 2 persons or alternatively nc = 3 persons mixtures, from an average of 32 bands in R1 and R2 DNA profiles (Table 2). The estimated upper d bound on the value of the probability of drop-out is, PrðDÞ 0:95 ¼ 0:16 and 0.38, respectively (Appendix I). From a preliminary assessment of evidence in this case, the first iteration of propositions is as follows. Proposition 1. Hp: V1 + V2 + S and Hd: V1 + V2 + U. However, examination of the DNA results suggested a possible alternative explanation. All of the alleles that could be attributed to victim two are shared with either victim one or the suspect. Therefore, the propositions could be modified as follows. Proposition 2. Hp: V1 + S and Hd: V1 + U. Now we condition upon a two person mixture. However, this would require five alleles to be explained as

contamination events (D18-16, D21-31.2, D2-23, D16-12, VWA-17). As Pr(C) = 0.05 per DNA profile, this would be unlikely. A more plausible explanation would be that DNA from three contributors was present, where one was unknown under Hp and Hd (i.e. transfer of DNA to the hammer from an unknown person could have occurred before the crime event). The absence of a DNA profile from V2 does not imply that she was not hit with the hammer, since transfer of DNA as a result of physical contact is dependent upon unquantifiable factors and is not a foregone conclusion [18]. Proposition 3. Hp: V1 + S + U and Hd: V1 + U1 + U2. For illustrative purposes only we also consider two separate, albeit highly improbable, propositions (since we believe that V2 DNA is absent), but it is interesting to determine the effect on the LR if V2 is substituted for V1. Proposition 4. Hp: V2 + S and Hd: V2 + U. Proposition 5. Hp: V2 + S + U1 and Hd: V2 + U1 + U2. Finally, to illustrate an unbalanced pair of propositions where Hp is anchored on V1 and S we evaluate Hd using V1 + U1 + U2—since nc is different under Hp and Hd, Pr(D) is conditioned on nc = 2 and 3, respectively. Proposition 6. Hp: V1 + S1 and Hd: V1 + U1 + U2. The probability of contamination was kept constant (Pr(C) = 0.05) for all propositions; with Pr(D) varied from 0.01 to 0.95 by 0.05 increments. LRs were calculated across all loci for each level of Pr(D) (Fig. 1). The highest LRs were calculated using Proposition 6 Hp: V + S and Hd: V1 + U1 + U2, followed by Proposition 2 Hp: V1 + S and Hd: V1 + U. However, we restate that neither is optimal for court reporting for the reasons outlined previously. Whereas the proposition V1 + S appeared to favour Hp the most, given the large number of unknown alleles that cannot be realistically explained by contamination, we advocate Hp: V1 + S + U as the simplest and most realistic prosecution proposition. Proposition 1: Hp: V1 + V2 + S and Hd: V1 + V2 + U and Proposition 3: V1 + S + U2 and Hd: V1 + U1 + U2 give LRs that are very similar. The substitution of V2 with U2 makes very little difference to the result, i.e. it does not assist the defence to argue whether V2 is present or whether an unknown person was present in the crime profile. The lowest LRs were calculated with Proposition 4: Hp: V2 + S and Hd: V2 + U. This result was not unexpected, as seven (out of twenty) of the alleles of victim two were not reproduced in any of the amplification replicate results—giving a much smaller numerator value.

P. Gill et al. / Forensic Science International 166 (2007) 128–138

133

Fig. 1. Casework example, log10 genotype likelihood ratios vs. probability of drop-out for each pair of propositions tested. The large striped arrows correspond d d to the x-axis estimate of probability of drop-out PrðDÞ 0:95 for nc = 2 and the large solid arrows estimate probability of drop-out PrðDÞ0:95 for nc = 3 (32 allele profile). A horizontal line to the y-axis gives an estimate of the log10 LR. For each line on the graph, the alternative prosecution and defence propositions are given in the format Hp/Hd.

Finally, Proposition 6: Hp: V1 + S and Hd: V1 + U1 + U2 gave the greatest LR, but as previously indicated; invoking multiple independent contaminant alleles is not particularly realistic and was therefore not advocated. Proposition 3 was preferred, whilst noting that Proposition 1 made very little difference with respect to the final LR at the predicted drop-out level Pr(D) = 0.38. The main purpose of this demonstration was to show how easy it is to rapidly evaluate any propositions required by the court. An important feature is that all calculations are relatively insensitive to Pr(D) since the fall in LR was small over the realistic range of Pr(D). 7.5. Comparison with the consensus model The consensus, or biological model results, evaluated Hp: V1 + V2 + S and Hd: V1 + V2 + U and the LR = Pr(EjHp)/ Pr(EjHd) = 1.55  106. This was conservative relative to all propositions tested except for the unrealistic pair of Propositions 4: Hp: V2 + S and Hd: V2 + U.

8. Discussion Whereas the contamination parameter is relatively straightforward to estimate from experimental observation of negative controls [4], the drop-out parameter is more problematic. Under the assumption that allelic drop-out is random [8] we currently estimate the distribution of this parameter from the number of alleles present in the DNA profile, relative to profiles randomly generated from a reference population database such as Caucasian. Different distributions result from different population databases—but the differences are minor (data not shown). It is currently impracticable to estimate multiple drop-out parameters (one for each potential contributor), consequently we effectively use an average (unweighted) value. It is informative to evaluate the effect of altering the drop-out parameter of individual loci comparing Hp: S + U and Hd: U1 + U2 (Table 5). Under Hp, S = ab and U = cd. To simplify calculations we evaluate a locus where alleles are either common ( p = 0.1) or rare ( p = 0.02). We have not considered

134

P. Gill et al. / Forensic Science International 166 (2007) 128–138

Table 5 LRs calculated for typical drop-out and contamination events at a single locus where Pr(allele) = 0.1 or 0.02, respectively, evaluating Hp: S + U and Hd: U1 + U2 Condition

No drop-out; no contamination 1 suspect allele dropped out 1 unknown allele dropped out Both suspect alleles dropped out 1 contamination event; no drop-out 1 contamination event; 1 suspect allele dropped out

Match probability of allele = 0.1

Match probability of allele = 0.02

Pr(D) = 0.1

Pr(D) = 0.5

Pr(D) = 0.9

Pr(D) = 0.1

Pr(D) = 0.5

Pr(D) = 0.9

8.3 0.4 9.1 0.21 5 0.03

8.3 0.98 2.5 0.21 5 0.24

8.3 1.16 0.3 0.21 5 1.3

208 3.44 77.5 0.17 125 0.14

208 4.3 10.7 0.17 125 1.2

208 4.42 1.2 0.17 125 6.7

the effect of F ST in these comparisons. Nevertheless, we illustrate that the following generalisations are useful when evaluating any locus: (a) If it is not necessary to invoke drop-out or contamination under Hp in order to explain S then the LR is constant because Pr(D) cancels out in the numerator and denominator. (b) If one S allele has dropped out then the evidence tends to be neutral, or favours Hd, dependent upon whether the remaining S allele is rare. (c) If both S alleles have dropped out, i.e. complete locus dropout under Hp then the evidence always favours Hd independent of Pr(D). (d) Similarly, Pr(D) cancels when a contamination event occurs provided both suspect alleles are present—the profile is type abcde. Hp is favoured. (e) If one contaminant band and one drop-out event has occurred under Hp, then the LR will favour Hd; the greater Pr(D), the greater the LR becomes. (f) Conversely, if an unknown allele is alleged to have dropped out under Hd, then this also reduces the LR—the greater Pr(D), the lower the LR becomes. The biggest effect occurs when Hp can only be explained if drop-out has occurred (e.g. the profile is abd) regardless of the value of Pr(D) chosen within the range 0.1 < Pr(D) < 0.9, the LR drops by an approximate order of magnitude within this range. In addition, the lower the Pr(D) the less likely it is that drop-out is a satisfactory explanation under Hp, and consequently the lower the LR becomes.

under Hd was of trivial consequence. This leads us to propose a possible new approach to assist in the evaluation of evidence. Reasonable (multiple) pairs of propositions can be selected in agreement with the court requirements. A minimum LRmin (the lowest LR calculated) can provide a base-line. It is worth noting that all propositions will have S in the numerator substituted by U in the denominator, i.e. we have shown that any differences between LRs are a result of secondary issues that relate to the number and conditioning of contributors to the crime stain evidence. If LR differences are trivial or bounded by LRmin, then the court may view that the peripheral issues are simply not relevant to the evidence, as it does not affect the primary consideration of whether the suspect contributed to the crime stain. If there are several alleles from an unknown source in a crime sample, then it is unlikely that these are explained by a contamination probability which is strictly only valid under the assumption that the contaminant alleles present are independent, and not from a single source. With Pr(C) = 0.05, on average, we would expect only one to two contaminant alleles. Consequently we recommend that profiles with three or more alleles that cannot be explained by the casework circumstances are always evaluated by invoking an addition unknown (U) contributor as the most reasonable explanation. The second recommendation is to use the Q designation with caution under Hp, since it always increases Pr(EjHp). Conversely, to maximise Pr(EjHd) it is reasonable to use Q if the alleles are at low level. 8.2. LoComatioN as a LR calculator for ‘conventional’ DNA profiles

8.1. General conclusions on forming propositions LoComatioN enables rapid evaluation of multiple propositions. Sometimes it is difficult to formulate propositions in casework because of uncertainties surrounding the casework circumstances. This is especially true for DNA profiles where the amount of DNA is limited. In addition, there may be ample opportunity for transfer of DNA to have occurred before the crime event. The case example described provided an opportunity to evaluate the effect of choosing different propositions for analysis. The profile was a mixture where it was unclear whether a victim’s DNA was present. We showed that the issue of whether V2 or U was the best explanation

LoComatioN can also be used to calculate LRs from conventional 28 cycle DNA profiles as well. There is a misconception that the low copy number definition applies only to elevated PCR cycle number. However, the defining feature of LCN is drop-out and drop-in. These phenomena also occur with 28 PCR cycles. Most laboratories have guidelines to indicate whether a given profile is sufficient for conventional interpretation (i.e. precluding allele drop-out). Many will report major/minor mixtures where the minor component is attributed to the suspect under Hp, but allele drop-out may be observed. All of the considerations described previously, also apply to low level DNA analysed using 28 PCR cycles.

P. Gill et al. / Forensic Science International 166 (2007) 128–138

If the alleles at a locus are above an experimentally defined threshold level (e.g. 150rfu) then allele drop-out is unlikely to occur. Under these conditions, Pr(D)  0 and consequently the Q designation is not relevant to the calculation of the LR. Under these conditions the theory used by LoComatioN converges to models previously described [19]—however, the advantage is that Pr(C) can be incorporated, multiple propositions can be evaluated, and furthermore the information from several replicates can be combined into one LR if necessary. Appendix A A.1. Simulation of the empirical likelihood for the probability of drop-out In the following simulations we consider the number of contributors, nc, and the probability of contamination Pr(C) to be fixed in advance. The goal of the simulations is to estimate the probability of observing x alleles at L loci given that the probability of drop-out is equal to D, Pr(D) = D. That is, we wish to estimate Pr(xjD, C, nc). Given that Pr(C) and nc are constant, this becomes Pr(xjD). The problem is that we do not know D. Therefore we use the data, x, to estimate D using maximum likelihood estimation. This quantity is called the likelihood of D and is denoted L(D). However, we do not know the likelihood function of D given x either, so we have constructed a simulation in order to estimate the likelihood function of D given x. As L(D) is estimated from simulation we call it the empirical likelihood of D. A.1.1. Simulation details There are three parts to the simulation. Firstly we must specify the value of D. Secondly, we must repeatedly generate nc random DNA profiles and combine them together subject to

135

drop-out. Finally we must consider that contamination may have occurred. Each iteration of the simulation (for a given value of D) will produce a random profile that could have resulted from the contribution of nc unrelated individuals profiles, and from this profile we can count the number of observed alleles, x. Note that because we are not considering quantitative information such as peak heights or areas, it is possible for allele masking to occur. For example, if nc = 2 and two random profiles are ab and bc, we will only observe abc in the resulting scene stain. Hence, even with no drop-out (Pr(D) = 0), it is possible to observe fewer than 2ncL alleles. The frequency with which different values of x occur for a given value of D is estimate of Pr(xjD). A.1.2. Simulation pseudo-code Descriptions of simulations are always problematic. For that reason, we describe out simulation in pseudo-code so that those who are interested may replicate the work. for D = 0.0, 0.01, 0.02, . . ., 0.90 let ˜ f ¼ ½0; . . . ; 0, where ˜ f is a vector of length (2nc + 1)L + 1 for i = 1, . . ., N Make the scene profile blank for j = 1, . . ., nc for l = 1, . . ., L Select two alleles at random, Al1, Al2 with probability pAlk , k = 1, 2 Generate two random uniform numbers, u1, u2  U[0, 1] If u1  D then add allele Al1 to the scene profile If u2  D then add allele Al2 to the scene profile for l = 1, . . ., L Generate a random uniform number, u  U[0, 1] If u Pr(C) add a random allele Al1, selected with probability pAl1 to the scene profile Record x, the total number of alleles observed Let f x = f x + 1 (the elements of ˜ f are labelled 0 to (2nc + 1)L) let Pr(xjD) = ( f x/N), x = 0, 1, . . ., (2nc + 1)L

where L is the number of loci in the multiplex (L = 10 for SGM+), N is the number of iterations per value of D. Increasing N will reduce the Monte Carlo sampling error in px. pAlk is the frequency of the kth allele at the lth locus in the population database. Note that usage above just means we select alleles randomly with probability proportional to their frequency in the database (population). The range of x is from 0 to (2n + 1)L because each individual can contribute at most two distinct peaks and furthermore we allow at least one contaminant allele per locus which may also be distinct. So when n = 2, there is a possibility that we will observe 0, . . ., 5 peaks and 0, . . ., 50 peaks over 10 loci.

Fig. 2. The likelihood surface for the probability of drop-out, given two contributors and Pr(C) = 0.05.

A.1.3. Simulation results Fig. 2 shows the likelihood surface for the probability of drop-out, given two contributors (nc = 2) and Pr(C) = 0.05. How is this used? This is best demonstrated by example. Consider the case in Section 5. A total of 32 alleles were observed across ten loci. Let us initially postulate that there were only two contributors to this profile. If x is constant, at 32, then the graph in Fig. 1 lets us answer the question ‘‘what is the most likely value for Pr(D) if x = 32?’’ We do this taking a

136

P. Gill et al. / Forensic Science International 166 (2007) 128–138

Fig. 3. Likelihood function for the probability of drop-out when x = 32 and nc = 2.

‘‘slice’’ of Fig. 1 along the line x = 32. This yields the graph in Fig. 3. From Fig. 3 we can see the maximum occurs when Pr(D) = 0. This means that 32 alleles are not uncommon when there is no drop-out and two contributors to the stain. However, we can see that it is also quite probable that we would observe 32 alleles even if Pr(D) = 0.2. Actually it is about 16 times less likely, but the point we wish to make is that it is not impossible to observe 32 alleles when Pr(D) = 0.2. Therefore, what we would like to do is put some sort of confidence bound on Pr(D). That is, we would choose a value D* so that 95% of intervals of the form [0, D*] would contain the true value. Although we use 95% in as an example throughout this paper there is no reason why a more stringent value (e.g. 99.9%) could not be used. To do this we need to estimate the cumulative distribution function (cdf) for the probability of drop-out given a certain value of x. We can change the likelihood function in Fig. 3 to a probability function by normalising it—i.e. making sure that the area under the curve sums to one (Fig. 4). In doing this, we are making the assumption that the probability of drop-out is a discrete random variable.1 In theory it is not, but in practice if we know the probability of drop-out to the nearest 1% (0.01) then this will be sufficient to calculate the LR without substantial bias to the defendant. Once we have the probability function for D, f(Djx), we can calculate the cumulative distribution function: FðDjxÞ ¼

d¼D X

f ðD ¼ djxÞ

d 2 f0;0:01;0:02;...g

The actual level of drop-out used in the LR calculations was taken from the 5th or 95th percentile of the cdf, dependent upon 1

And we are implicitly placing a uniform prior on it as well. Technically the normalization of the likelihood is a Bayesian operation, hence the interpretation of the resulting intervals are correct in a Bayesian sense.

Fig. 4. The cumulative distribution function (cdf) F(Djx = 32) for a profile with 32 alleles. The solid line is the cdf for D assuming that there are three (nc = 3) contributors to this mixture, whereas the dashed line is the cdf for D assuming that there are two (nc = 2) contributors. The y-axis tells us the probability that D is smaller than the value on the x-axis. For example, if a vertical line from the xaxis is drawn at the point 0.16 to where it hits the dashed line, and a horizontal line to the y-axis, it hits at about 0.95. We interpret this as ‘‘assuming only two people contributed to this mix, we are 95% sure that the true value of Pr(D) is less than 0.16.

the level that minimised the LR—in practice this is usually the 95th percentile. Mathematically we evaluate qa = F 1(a) where F 1(a) inverse cumulativeRdistribution function is given x by finding the value x such that 1 f ðtÞ dt ¼ a. a = 0.05 for the 5th percentile and a = 0.95 for the 95th percentile. In practice we approximate the cdf as a piecewise linear function. We find two points q1 and q2, such that F(q1) < a < F(q2) and we return F 1 ðaÞ  wq1 þ ð1  wÞq2 where w ¼ ða  Fðq1 ÞÞ=ðFðq2 Þ  Fðq1 ÞÞ. In our example this yields values of 0.16 and 0.38. Appendix B. A more detailed example of LoComatioN principles In LoComatioN [12] the Q allele designation enables probabilistic evaluation of all possible allelic combinations, including those that could be explained if drop-out and contamination had happened. From the casework example, we evaluate all possible allele propositions for each locus in turn. For example for the case stain evidence (E) at the D3 locus we have two identical results: R1 = R2 = 14,16. The suspect, S, has genotype 14,16 and the victim V, has genotype 16,16. The propositions under consideration are:  Hp: the victim, suspect and one unknown unrelated contributor are the only people who have contributed to this stain (V + S + U);  Hd: the victim and two unknown unrelated contributors are the only people who have contributed to this stain (V + U1 + U2).

P. Gill et al. / Forensic Science International 166 (2007) 128–138

137

Table 6 Illustration of probabilistic principles employed to formulate the probabilities under Hp Proposed contributing genotypes V + S + U

Pr(R1 = 14,16jMj)

Pr(R2 = 14,16jMj)

Pr(Mj)

Product

16,16 + 14,16 + 14,14

No drop-out, no ¯ ¯ 6 PrðCÞ contamination PrðDÞ 6 ¯ ¯ PrðCÞ PrðDÞ ¯ ¯ 6 PrðCÞ PrðDÞ No drop-out, drop-out and ¯ ¯ 5 PrðDÞPrðCÞ contamination PrðDÞ 5 ¯ ¯ PrðDÞPrðCÞ PrðDÞ ¯ ¯ 4 PrðDÞPrðCÞ PrðDÞ

No drop-out, no ¯ ¯ 6 PrðCÞ contamination PrðDÞ 6 ¯ ¯ PrðCÞ PrðDÞ ¯ ¯ 6 PrðCÞ PrðDÞ No drop-out, drop-out and ¯ ¯ 5 PrðDÞPrðCÞ contamination PrðDÞ 5 ¯ ¯ PrðDÞPrðCÞ PrðDÞ ¯ ¯ 4 PrðDÞPrðCÞ PrðDÞ

p314 p316

¯ 2 p314 p316 ¯ 12 PrðCÞ PrðDÞ

2 p214 p416 p14 p516 2 p214 p316 pQ

¯ 2 p214 p416 ¯ 12 PrðCÞ 2PrðDÞ 12 ¯ ¯ PrðDÞ PrðCÞ2 p14 p516 ¯ 2 p214 p316 pQ ¯ 10 PrðDÞ2 PrðCÞ 2PrðDÞ

2 p14 p416 pQ p14 p316 p2Q

¯ 2 p14 p416 pQ ¯ 10 PrðDÞ2 PrðCÞ 2PrðDÞ 8 4 ¯ 2 p14 p316 p2Q ¯ PrðDÞ PrðCÞ PrðDÞ

16,16 + 14,16 + 14,16 16,16 + 14,16 + 16,16 16,16 + 14.16 + 14,Q 16,16 + 14,16 + 16,Q 16,16 + 14,16 + Q,Q

The numerator is then calculated by summing the entire product column, using the total law of probability.

Evaluation of the probability of the evidence under Hp is straight-forward – the unknown contributor, U, is allowed to have a genotype formed by any combination of alleles 14, 16 and Q – allowing for the possibility of drop-out to be considered. Hence, the genotypes considered for the unknown contributor, under Hp, would be: 14,14; 14,16; 16,16; 14,Q; 16,Q; Q,Q.

In order to illustrate the probabilistic principles employed in the software, the calculations have been formulated for the Hp alternatives in Table 6. The Hd calculations proceed in a similar fashion, however under Hd there are two unknown contributors, making the list of possible alternative genotypes for U1 and U2 a great deal longer, see Fig. 5 for allele combination listings. The following

Fig. 5. LoComatioN screen-shot showing some of the allelic combinations to be considered under Hp: V + S + U and Hd: V + U1 + U2 from a casework example (Table 2) LR = Pr(EjHp)/Pr(EjHd). Under Hd, all potential genotypes from U1 + U2 contributors are considered.

138

P. Gill et al. / Forensic Science International 166 (2007) 128–138

Table 7 Expansion of the first four rows of Fig. 5, to illustrate probabilistic principles employed to formulate probabilities under Hd Proposed contributing genotypes V + U1 + U2

Pr(R1 = 14,16jMj)

Pr(R2 = 14,16jMj)

Pr(Mj)

Product

16,16 + 14,14 + 14,14

No drop-out, no ¯ ¯ 6 PrðCÞ contamination PrðDÞ 6 ¯ ¯ PrðCÞ PrðDÞ No drop-out, drop-out and ¯ ¯ 5 PrðDÞPrðCÞ no contamination PrðDÞ 4 2 ¯ ¯ PrðDÞ PrðCÞ PrðDÞ

No drop-out, no ¯ ¯ 6 PrðCÞ contamination PrðDÞ 6 ¯ ¯ PrðCÞ PrðDÞ No drop-out, drop-out and ¯ ¯ 5 PrðDÞPrðCÞ no contamination PrðDÞ 4 2 ¯ ¯ PrðDÞ PrðCÞ PrðDÞ

p514 p316

¯ 2 p514 p316 ¯ 12 PrðCÞ PrðDÞ

4 p414 p416 4 p414 p316 pQ

¯ 2 p414 p416 ¯ 12 PrðCÞ 4PrðDÞ 10 ¯ 2 p414 p316 pQ ¯ 4PrðDÞ PrðDÞ2 PrðCÞ

6 p314 p316 p2Q

¯ 2 p314 p316 p2Q ¯ 8 PrðDÞ4 PrðCÞ 6PrðDÞ

16,16 + 14,14 + 14,16 16,16 + 14,14 + 14,Q 16,16 + 14,14 + Q,Q

Table 7 has been included in order to demonstrate that the principles applied to Hp, also apply to Hd. References [1] I. Findlay, A. Taylor, P. Quirke, R. Frazier, A. Urquhart, DNA fingerprinting from single cells, Nature 389 (1997) 555–556. [2] P. Gill, R. Sparkes, C. Kimpton, Development of guidelines to designate alleles using an STR multiplex system, Forens. Sci. Int. 89 (1997) 185– 197. [3] J.P. Whitaker, E.A. Cotton, P. Gill, A comparison of the characteristics of profiles produced with the AMPFlSTR SGM Plus multiplex system for both standard and low copy number (LCN) STR DNA analysis, Forens. Sci. Int. 123 (2001) 215–223. [4] P. Gill, A. Kirkham, Development of a simulation model to assess the impact of contamination in casework using STRs, J. Forens. Sci. 49 (2004) 485–491. [5] M. Bill, P. Gill, J. Curran, T. Clayton, R. Pinchin, M. Healy, J. Buckleton, PENDULUM—a guideline based approach to the interpretation of STR mixtures, Forens. Sci. Int. 148 (2004) 181–189. [6] T.M. Clayton, J.P. Whitaker, R. Sparkes, P. Gill, Analysis and interpretation of mixed forensic stains using DNA STR profiling, Forens. Sci. Int. 91 (1998) 55–70. [7] P. Gill, R. Sparkes, R. Pinchin, T. Clayton, J. Whitaker, J. Buckleton, Interpreting simple STR mixtures using allele peak areas, Forens. Sci. Int. 91 (1998) 41–53. [8] P. Gill, J. Curran, K. Elliot, A graphical simulation model of the entire DNA process associated with the analysis of short tandem repeat loci, Nucleic Acids Res. 33 (2005) 632–643.

[9] P. Gill, J. Whitaker, C. Flaxman, N. Brown, J. Buckleton, An investigation of the rigor of interpretation rules for STRs derived from less than 100 pg of DNA, Forens. Sci. Int. 112 (2000) 17–40. [10] J. Buckleton, P. Gill, Low copy number, in: J. Buckleton, C.M. Triggs, J.S. Walsh (Eds.), Forensic DNA Evidence Interpretation, CRC Press, 2005, pp. 275–297. [11] P. Taberlet, S. Griffin, B. Goossens, S. Questiau, V. Manceau, N. Escaravage, L.P. Waits, J. Bouvet, Reliable genotyping of samples with very low DNA quantities using PCR, Nucleic Acids Res. 24 (1996) 3189–3194. [12] J.M. Curran, P. Gill, M.R. Bill, Interpretation of repeat measurement DNA evidence allowing for multiple contributors and population substructure, Forens. Sci. Int. 148 (2005) 47–53. [13] C.H. Brenner, R. Fimmers, M.P. Baur, Likelihood ratios for mixed stains when the number of donors cannot be agreed, Int. J. Legal Med. 109 (1996) 218–219. [14] B.S. Weir, DNA statistics in the Simpson matter, Nat. Genet. 11 (1995) 365–368. [15] J. Buckleton, J.M. Curran, P. Gill, Towards understanding the effect of uncertainty in the number of contributors to DNA stains, Forens. Sci. Int., in press. [16] I.W. Evett, G. Jackson, J.A. Lambert, More on the hierarchy of propositions: exploring the distinction between explanations and propositions, Sci. Justice 40 (2000) 3–10. [17] R. Cook, I.W. Evett, G. Jackson, P.J. Jones, J.A. Lambert, A model for case assessment and interpretation, Sci. Justice 38 (1998) 151–156. [18] A. Lowe, C. Murray, J. Whitaker, G. Tully, P. Gill, The propensity of individuals to deposit DNA and secondary transfer of low level DNA from individuals to inert surfaces, Forens. Sci. Int. 129 (2002) 25–34. [19] I.W. Evett, C. Buffery, G. Willott, D. Stoney, A guide to interpreting single locus profiles of DNA mixtures in forensic cases, J. Forens. Sci. Soc. 31 (1991) 41–47.

Forensic Science International 166 (2007) 139–144 www.elsevier.com/locate/forsciint

Morphine and 6-acetylmorphine concentrations in blood, brain, spinal cord, bone marrow and bone after lethal acute or chronic diacetylmorphine administration to mice Emmanuelle Guillot a, Philippe de Mazancourt a, Michel Durigon b, Jean-Claude Alvarez a,* a

Laboratoire de Pharmacologie–Toxicologie, Centre Hospitalier Universitaire Raymond Poincare´, AP-HP, 104 Boulevard R. Poincare´, 92380 Garches, France b Service de Me´decine Le´gale, Centre Hospitalier Universitaire Raymond Poincare´, AP-HP, 104 Boulevard R. Poincare´, 92380 Garches, France Received 3 February 2006; received in revised form 23 March 2006; accepted 23 March 2006 Available online 24 May 2006

Abstract The aim of this study was to evaluate postmortem incorporation of opiates in bone and bone marrow after diacetylmorphine (heroin) administration to mice. Mice were given acute (lethal dose of 300 mg/kg) or chronic (10 and 20 mg/kg/24 h for 20 days) intraperitoneal administration of diacetylmorphine. The two metabolites of diacetylmorphine, 6-acetylmorphine (6-AM) and morphine, were extracted from whole blood, brain, spinal cord, bone marrow and bone (after hydrolysis) using a liquid/liquid method. Quantification was performed by gas chromatography–mass spectrometry (GC/MS). Results showed that after acute administration, opiates were present in all studied tissues. Morphine concentrations appeared to be higher than those of 6-AM in blood (52.4 mg/mL versus 27.7 mg/mL, n = 12), bone marrow (87.8 ng/mg versus 8.9 ng/mg, n = 6) and bone (0.85 ng/mg versus 0.43 ng/mg, n = 6), but 6-AM concentrations were higher than those of morphine in brain (14.0 ng/mg versus 7.4 ng/mg, n = 12) and spinal cord (27.8 ng/mg versus 20.8 ng/mg, n = 12). No correlation was found for both compounds between blood concentrations and either brain, spinal cord, bone or bone marrow concentrations while a significant one was found between brain and spinal cord concentrations either for morphine (r = 0.89, n = 12, p < 0.001) or 6-AM (r = 0.93, n = 12, p < 0.001), the concentration being higher in spinal cord than in brain. When bones were stored for 2 months, only 6-AM remained in bone marrow but not in bone. After chronic administration, mice being sacrificed by cervical dislocation 24 h after the last injection, no opiate was detected in any studied tissues. Further studies are required, in particular in human bones, but these results seem to show that 6-AM could be detect in bone marrow several weeks after the death and could be an alternative tissue for forensic toxicologist to detect a fatal diacetylmorphine overdose, even if no correlation between blood and bone marrow was observed. On the other hand, neither bone tissue nor bone marrow will allow the confirmation of a chronic diacetylmorphine use. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Bone; Bone marrow; Diacetylmorphine; Overdose; Toxicology

1. Introduction When bodies are extremely putrefied, specimens such as blood or urine are not available for forensic toxicologists to detect the presence of drugs or toxicological substances. Many studies have been carried out for a few years to find alternative tissues, which could remain available for detection of drugs a

* Corresponding author. Tel.: +33 1 47 10 79 38; fax: +33 1 47 10 79 23. E-mail address: [email protected] (J.-C. Alvarez). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.03.029

long time after death. Hair samples [1,2], nails [3] or teeth [4] have already been used for postmortem detection of opiates, as well as bone and bone marrow [5]. Diacetylmorphine is frequently involved in fatal overdose cases. It undergoes rapid esterase hydrolysis to 6-acetylmorphine (6-AM), which is further deacetylated to morphine [6]. Morphine is conjugated to morphine-3 and 6-glucuronide in humans, and preferentially to morphine-3-glucuronide in rats and mice [7]. Bone marrow has a rich vascular supply and a lipid matrix that may allow drug diffusion and incorporation from blood [8].

140

E. Guillot et al. / Forensic Science International 166 (2007) 139–144

Moreover, it is encased in bone thus protecting it from possible contamination [9] or degradation. Several drugs including amitriptyline [8], metamphetamine and amphetamine [10], triazolam [11] and bromisovalum [12] have been detected and quantified in human bone marrow. Two conditions are required to ensure a diagnosis of fatal drug intoxication on bone marrow samples. First, a correlation between blood and bone marrow concentrations must be established, and secondly, bone marrow concentrations have to remain stable after death until the analysis is performed. A linear relationship between blood and bone marrow levels was found for several drugs by Winek et al. [13–15]. Moreover, metamphetamine and amphetamine concentrations in bone marrow collected from bones stored for 2 years in open air showed no drug degradation over this period [16]. However, a literature review showed a lack of available data about opiates. Drug detection in bone has only been studied in a few publications [5,17–19]. However, as with hair, drugs may remain trapped in bone for a long time. If so, it can be useful in determining a regular diacetylmorphine intake. Incorporation of benzodiazepines was reported in mice bone after chronic exposure, but not after acute administration [19]. In another study, determination of morphine in human postmortem bone and bone marrow was achieved, while 6-AM was not detected, due to bone demineralization procedure according to the authors [5]. Our preliminary hypotheses were that bone could be representative of chronic diacetylmorphine use, while bone marrow could be representative of diacetylmorphine overdose. So, the purpose of this study was to investigate the incorporation of opiates in bone and bone marrow after lethal acute or chronic diacetylmorphine injections to mice. In order to investigate any correlation between these tissues and blood or central nervous system, we also investigate the concentrations of morphine and 6-AM in blood, brain and spinal cord. Finally, we investigate the stability of the compounds in bone and bone marrow after 2 months. 2. Materials and methods 2.1. Experimental design Experiments were performed on 28 Swiss mice weighing 15 g, housed seven per cage under 12-h light:12-h dark cycle in a constant ambient temperature of 24  1 8C, with a free access to food and water (the principles of French laboratory animal care were followed). Animals received intraperitoneal injections of either saline (one control mouse per group) or diacetylmorphine hydrochloride (Euromedex, Mundolsheim, France) dissolved in saline (six mice per group). Two diacetylmorphine administration protocols were used in this study: - Acute administration: Two groups of six mice, respectively, received a lethal diacetylmorphine injection of 300 mg/kg. In the first group of animals, bone marrow samples were collected immediately. All samples (blood, brain, spinal cord,

marrow and pieces of bone) were frozen at 20 8C until analysis. In the second group, blood, brain and spinal cord were frozen immediately, but bones were stored in a jar containing 150 mL of non-sterile soil found in the forest. The bones were put on the soil and stored at room temperature. Two months later, dried bone marrow was collected and was frozen with bones. - Chronic administration: Two groups of six mice, respectively, received chronic diacetylmorphine injections of 10 or 20 mg/ kg every 24 h for 20 days. The animals were sacrificed on the 21st day, 24 h after the last administration, by cervical dislocation. Blood, brain, spinal cord, bones and bone marrow were collected and frozen at 20 8C until analysis. 2.2. Extraction procedures The gas chromatography–mass spectrometry (GC–MS) technique used in this study was adapted for each specimen from the procedure recommended by French Society of Analytical Toxicology [20] for the analysis of opiates in blood. 2.2.1. Chemicals Stock methanolic solutions of morphine (1.0 mg/mL), 6-AM (1.0 mg/mL), morphine-d3 (0.1 mg/mL) and 6-AM-d3 (0.1 mg/mL) were purchased from Promochem (Molsheim, France). Bistrimethylsilyltrifluoroacetamide (BSTFA)/trimethylchlorosilane (TMCS) (99/1, v/v) was purchased from Supelco (Bellefonte, USA). Other chemicals used were of chromatographic or analytical grade. 2.2.2. Extraction procedure in blood, brain, spinal cord and bone marrow Weighed bone marrow, brain and spinal cord were diluted and homogenized with an Ultraturax in 0.5 mL of deionized water for bone marrow and spinal cord and 4 mL for brain. To 500 mL of whole blood, brain, spinal cord or bone marrow homogenates were added 12.5 mL of 10 mg/L methanolic solution of internal standards (IS) (morphine-d3, 6-AM-d3), 1 mL of 2 M phosphate buffer pH 8.4 and 8 mL of chloroform/isopropanol/n-heptane: 50/17/33 (v/v/v). The preparation was shaken for 15 min and centrifuged at 3500  g for 5 min. The aqueous layer was discarded and the solvent layer was extracted by 5 mL of 0.2 M HCl. The aqueous layer obtained after agitation and centrifugation was neutralized by 1 mL of 1 M NaOH and re-extracted by 2 mL of phosphate buffer pH 8.4 and 5 mL chloroform. After agitation and centrifugation, the organic phase was evaporated under a nitrogen stream. Finally, the extract was derivatized by addition of 20 mL of BSTFA/TMCS and 30 mL of ethylacetate and heated at 70 8C for 30 min. Two microliters of the resulting solution were injected into the GC column. 2.2.3. Extraction procedure in bone Bone pieces were cleaned and boiled for 30 min to ensure that all bone marrow and muscle tissues had been removed. They were then ground in a mortar. To 50 mg of bone powder were added 25 mL of 10 mg/L IS methanolic solution and 1 mL

E. Guillot et al. / Forensic Science International 166 (2007) 139–144

of 0.1 M HCl. The mixture was incubated at 55 8C for 12 h in order to perform bone hydrolysis. This hot acid hydrolysis was adapted from a method for the determination of hair opiates described by Kintz and Mangin [21]. One milliliter of 0.1 M NaOH was then added and the mixture was extracted using the procedure as described above. 2.3. Apparatus The GC/MS analysis was performed on a ThermoFinnigan GC 8060 Fisons equipped with a ThermoFinnigan mass spectrometer Automass II (electron impact mode). The capillary column used was a PTE 5 (30 m  0.25 mm i.d., 0.25 mm film thickness, Supelco, Bellefonte, USA). Samples were injected in the splitless mode at 250 8C. Temperature was programmed to rise from 90 8C (0.5 min) to 220 8C (30 8C/min, 3 min hold), and then to 290 8C (15 8C/min, 1 min hold). Transfer line and ion source temperatures were, respectively, 300 and 200 8C. Flow rate of the carrier gas (helium) was 2.1 mL/min. The ionization energy was 70 eV. The detection was in single ion monitoring mode at m/z 429 (quantification ion), 236 and 414 (confirmation ions) for morphine, and 399 (quantification ion), 340 and 287 (confirmation ions) for 6-AM. Ions at m/z 432, 239 and 417, and m/z 402, 343 and 290 were used for morphine-d3 and 6-AM-d3, respectively.

141

the peak-area of the compounds and the peak-area of their deuterated internal standards. The precision and accuracy of the method were carried out over 3 days in blood. Each day, two calibration curves with six determinations of three quality controls (20, 125 and 500 ng/ mL) were analyzed. The values obtained were analyzed using variance analysis (ANOVA), which separated the intradayassay and interday-assay standard deviations and consequently the corresponding coefficients of variation (CV). The intradayassay CV took into account the variability of the six replicates each day for 3 days and the interday-assay CV the variability of the days of analysis. The accuracy was determined by comparing the mean calculated concentration with the spiked target concentration of the quality control samples. The limit of detection (LOD) was defined as the lowest concentration of the compounds that can be detected with a signal-to-noise ratio greater than 5:1. The limit of quantification (LOQ) was defined as the lowest concentration of the analytes that can be quantified with both an accuracy of 10% of the spiked value and a coefficient of variation (CV) 20%. Extraction recoveries were estimated at 200 ng/mL (n = 6), and selectivity of the method was evaluated by analyzing blank blood and tissues (mice controls) by the described method. 2.5. Statistical analysis All samples were analyzed only once, because of the small amounts of samples collected from mice. Statistical analysis was performed using correlation test and unpaired Student’s t-test.

2.4. Method validation Two working solutions (10 and 1.0 mg/L) containing morphine and 6-AM were prepared by appropriate dilutions of stock solution in methanol. The working solution of morphine-d3 and 6-AM-d3 (IS) was prepared in methanol at 10 mg/L. All stock and working solutions were stored at 20 8C for a maximum of 3 months and 1 month, respectively. For linearity study, calibration curves were prepared by spiking six drug-free whole bloods (0.5 mL) with appropriate volumes of working solutions containing morphine and 6-AM (range 10–1000 ng/mL), or by spiking 50 mg of blank bone powder (range 0.25–20 ng/mg). Brain, spinal cord and bone marrow concentrations were determined with blood calibration curves. Quantification was performed by calculating the ratio between

3. Results Calibration curves were found to be linear in blood within the range 10–1000 ng/mL, with correlation coefficients being 0.995 for morphine and 0.992 for 6-AM. Linearity was also established in bone within the range 0.25–20 ng/mg. The LOD for both compounds were set at 3 ng/mL for blood analysis and 0.1 ng/mg for bone analysis, and the LOQ were set at 10 ng/mL and 0.25 ng/mg in blood and bone, respectively. The results of intra and inter-assay precision and accuracy of the method are presented in Table 1. The intra-assay coefficients of variation (CVs) were less than 6%, and the inter-assay coefficients of

Table 1 Intra- and inter-assay precision and accuracy of the method in blood Added (ng/mL)

Mean concentration found (ng/mL)

Accuracy (%)

Intra-assay CV (%)

Inter-assay CV (%)

Morphine 20 125 500

20.65 124.64 495.05

103.2 99.7 99.0

5.78 1.81 2.98

10.88 3.59 5.07

6-AM 20 125 500

20.13 122.73 492.62

100.7 98.2 98.5

4.82 4.20 2.98

10.87 8.85 4.08

Each day during 3 days, two calibration curves with six determinations of the three quality control samples (20, 125 and 500 ng/mL) were analyzed. The values obtained were analyzed using variance analysis (ANOVA).

142

E. Guillot et al. / Forensic Science International 166 (2007) 139–144

Table 2 Morphine and 6-AM concentrations in blood (samples were diluted 1:100), brain, spinal cord, bone marrow and bone after acute lethal diacetylmorphine administration Mouse

1 2 3 4 5 6

Blood (mg/mL)

Brain (ng/mg)

Spinal cord (ng/mg)

Bone marrow (ng/mg)

Bone (ng/mg)

Morphine

6-AM

Morphine

6-AM

Morphine

6-AM

Morphine

6-AM

Morphine

6-AM

19.8 10.8 40.3 17.6 88.0 44.8

7.1 8.1 11.4 6.7 16.7 36.7

2.7 3.2 12 9.3 8.9 17.3

2.7 1.3 13 7.9 16 12

1.8 3.5 17.7 13 17.7 27.5

69 34 220 58 112 34

4.95 5.5 24 2.4 10 6.8

1.9 0.4 1.2 0.4 0.83 0.43

0.30 0.26 0.60 0.38 0.72 0.30

6 0.85  0.24

6 0.43  0.08



– –

– –

2.4 0.64 7.3 4.4 5.6 6.2

n Mean  S.E.M.

6 87.8  28.9

7 8 9 10 11 12

83.7 86.3 58.9 50.6 68.4 59.7

57.0 29.8 33.0 56.0 18.9 50.4

14 15.5 4.9 16.2 7 4.1

14.5 30 8.2 34.9 12.4 14.5

46.5 52 23.4 32.6 26.2 16.5

20.7 71 29.8 61.6 33.5 35.3

N.D. N.D. N.D. N.D. N.D. N.D.

n Mean  S.E.M.

12 52.4  7.7

12 27.7  5.5

12 7.4  1.5

12 14.0  2.8

12 20.8  4.7

12 27.8  6.1

– –

6 8.9  3.2 6.3 0.7 1.2 2.6 0.3 4.2 6 2.6  1.0

For mice 1–6: blood, brain, spinal cord, bone marrow and bone were frozen immediately after death. For mice 7–12: blood, brain and spinal cord were frozen immediately after death. Bones were stored for 2 months. Dried bone marrow and bones were then frozen and analyzed together with blood, brain and spinal cord. N.D.: not detected.
variation were less than 11% at all concentrations tested. The recovery from spiked blood and bone was determined to be 96% for morphine and 89 and 80% for 6-AM, respectively. Specificity of the method was demonstrated since no signal was observed in all blank samples at the retention times of each compound with specific single ion monitoring detection. Table 2 presents blood, brain, spinal cord, bone marrow and bone morphine or 6-AM concentrations after lethal injection of diacetylmorphine. Blood concentrations were calculated in 100 diluted samples, in order to be in the calibrated range of the method. Time elapsed between injection and death ranged from 3 to 8 min. Mean blood, brain and spinal cord concentrations were given for 12 mice. Mean bone marrow and bone concentrations for mice 1–6 correspond to sample immediately frozen after death. For mice 7–12, bones were stored for 2 months. Morphine concentrations appeared to be higher than 6-AM concentrations in blood (52.4  7.7 mg/mL versus 27.7  5.5 mg/mL, n = 12), bone marrow (87.8  28.9 ng/mg versus 8.9  3.2 ng/mg, n = 6) and bone (0.85  0.24 ng/mg versus 0.43  0.08 ng/mg, n = 6). On the other hand, 6-AM concentrations are higher than those of morphine in brain (14.0  2.8 ng/mg versus 7.4  1.5 ng/mg, n = 12) and spinal cord (27.8  6.1 ng/mg versus 20.8  4.7 ng/mg, n = 12). No correlation was found between blood and brain or spinal cord (n = 12), or between blood and bone marrow or bone frozen immediately after sacrifice (n = 6). On the other hand, a significant correlation was observed between brain and spinal cord concentrations either for morphine (r = 0.895, n = 12, p < 0.001) or 6-AM (r = 0.926, n = 12, p < 0.001). After 2 months of storage, the average loss of 6-AM in bone marrow was 71%, while a complete loss of morphine was

observed. In bone, morphine and 6-AM were found at concentrations below the limit of quantification for both compounds. In the chronic study, either for the 10 mg/kg or 20 mg/kg dosage group, no opiate was detected in blood, brain, spinal cord, bone marrow or bone tissues collected 24 h after the last administration of diacetylmorphine. In all experiments, neither morphine nor 6-AM was detected in blood and tissues of controls mice. 4. Discussion Diacetylmorphine overdoses are frequently encountered in forensic cases. In cases where no soft tissue remains, bone and bone marrow could be one of the best materials in order to identify drugs, but no data are available about incorporation of opiates in these tissues. Our study used a GC/MS method to investigate opiates incorporation in mice bone and bone marrow. This method appeared to be accurate, reproducible and adequately sensitive for the range of concentrations studied in spite of low concentrations in mice bone, which were sometimes close to the limit of quantification. Preparation of bone pieces required attention because samples had to be cleaned from overlying muscle and removal of the bone marrow. A technique based on rinsing with deionized water as proposed by McIntyre et al. [17] was tested, but residues of marrow were visibly remaining on bone. With the procedure used in our design (bones boiled for 30 min), the pieces of bone seemed to be completely cleaned although it was difficult to ensure that all connective tissues had been removed. As 6-AM was known to be unstable [6,22], we have preliminary tested by incubating at 100 8C for 30 min

E. Guillot et al. / Forensic Science International 166 (2007) 139–144

whether this compound was deacetylated to morphine. Only a small amount of 6-AM was lost under these conditions (less than 10%). Similarly, several procedures for bone hydrolysis had been previously published (drug solubilization into hot methanol [17], bone demineralization by HNO3 solution [5]). The method we used here (hot acid hydrolysis) was adapted from a Kintz and Mangin [21] method for the determination of hair opiates. It optimized the balance between drug liberation and 6-AM hydrolysis. We observed that recovery of 6-AM was 80% after this hydrolysis, instead of 89% in whole blood. However, the use of deuterated internal standards introduced before hydrolysis, which have the same chemical properties than the analytes, compensated for any degradation of 6-AM and morphine during sample preparation. We used in our design a fatal overdose of 300 mg/kg, which is far away from those used in human overdose. However, preliminary assays had shown that lower doses did not allow detecting morphine and 6-AM in bone or bone marrow, and did not lead to the fatal effect in mice. As the initial purpose of this study was to evaluate the stability of these compounds in these tissues after lethal overdose, we decided to administrate higher doses of diacetylmorphine to the mice. As shown in Table 2, a large variation in all concentrations between animals of the same group was observed. This can probably be explained by the inter-individual variations in animal. Death occurred between 3 and 8 min after diacetylmorphine injection. To minimize this variability, mice should have been sacrificed at a same fixed time after the lethal injection. But in our design, waiting for the death of the mice was preferred in order to simulate a human diacetylmorphine overdose. We observed that 6-AM concentrations were higher than those of morphine only in brain and spinal cord, which are tissues rich in lipids. 6-AM is a more lipophilic and lipid soluble compound than morphine and has been shown to cross the blood–brain barrier more rapidly than morphine in mice [23]. Moreover, in brain, no metabolism of 6-AM in morphine may occur [24], contrary to what happens rapidly in blood according to the short half-life time of 6-AM [7,24]. A high diffusion from blood to bone marrow was observed especially for morphine after acute administration, even if no relationship was established between blood and bone marrow concentrations. Winek et al. found such correlation for many other drugs [9,13–15]. However, in these studies, Winek et al. did not use lethal doses. After a lethal diacetylmorphine dose, death may occur before that equilibrium was reached between blood and bone marrow concentrations, which could explain the absence of correlation. We only observed a strong correlation between brain and spinal cord, but this correlation was expected, since these two tissues are closely linked and are both located behind the blood–brain barrier. Contrary to what we hypothesized, morphine and 6-AM were present in bone after acute injection but not after chronic administration, which could be the probable reflection of the well vascularised status of this tissue and the absence of incorporation in this matrix. These results are not in accordance with those of Gorczynski and Melbye [19].

143

In this study, the authors detected benzodiazepines in bone after chronic administration and not after acute injection. However, this study was dealing with different compounds and a different time course of administration: animals were sacrificed 16 h after the last dose of chronic administration, whereas the benzodiazepines were still present in blood, brain, and bone marrow. Consequently, the presence of benzodiazepines in bone could be only due to the last injection and not to a chronic incorporation of the compounds in bone. In our design, we decided to sacrifice the mice 24 h after the last injection because no opiates were detectable in blood at this time. We decided to store the bones for 2 months in order to evaluate the stability of both compounds in this matrix and in bone marrow, because 2 months is enough to achieve skeleton when a body is left in the forest for example. No data about drug stability in animal bones were available in literature. Raikos et al. [5] analyzed immediately after death by FPIA and GC-FID bone and bone marrow of a human fatally poisoned diacetylmorphine addict, then a piece of the bone was buried for 1 year. No bone marrow was left in the piece of bone since both sides of bone were open during burial. The authors found morphine in fresh bone marrow and bone, in accordance with our results, and observed a loss of 54.4% of morphine concentration in buried bone. 6-AM was not found, probably due, according to the authors, to its hydrolysis to morphine during the extraction procedure. Consequently, the presence of morphine in buried bone could be due only to hydrolysis of 6-AM in morphine. This study provided any information neither about 6-AM presence or stability nor about morphine stability in bone marrow of buried bones. Our results showed that free morphine was totally lost in bone marrow when bones were stored for 2 months. However, the absence of free morphine did not allow concluding in the absence of morphine-glucuronides. Unfortunately, the low quantity available of bone marrow did not allow measuring morphine after cleavage of the morphine conjugate, which could lead to highlight presence of morphine if its conjugates were still present. Future studies should include total morphine and not only free morphine. On the other hand, 6-AM was still present, even if an important loss of 71% was observed. The greater stability of 6-AM than morphine could be explained by the possible formation of a stable complex of two 6-AM molecules linked by two water molecules, as previously shown in vitro [25]. Two conditions seem to be required to ensure a diagnosis of fatal drug intoxication on bone marrow samples. First, a correlation between blood and bone marrow concentrations may exist, and secondly, bone marrow concentrations have to remain stable after death until the analysis is performed. Both conditions were not established here. However, the highlight of the remaining 6-AM in bone marrow several weeks after the death could be sufficient for the use of this tissue as an alternative one matrix for forensic toxicologist to detect a fatal diacetylmorphine overdose, showing the usefulness of bone with respect to hair in exhumation cases, since hair might be used as a surrogate to blood to assess only chronic use.

144

E. Guillot et al. / Forensic Science International 166 (2007) 139–144

5. Conclusion Further studies are required to ensure the stability of opiates in human bone and bone marrow, which can be different from what was observed in little mice bone. But the present results seem to show that bone marrow could be suited to detect a fatal diacetylmorphine overdose, especially by the presence of 6AM. The time-period over which the reduction of the 6-AM took place has to be evaluated in human bone. On the other hand, neither bone tissue nor bone marrow will allow the confirmation of a regular diacetylmorphine intake. References [1] F. Tagliaro, Z. De Battisti, F.P. Smith, M. Marigo, Death from diacetylmorphine overdose: findings from hair analysis, Lancet 351 (1998) 1923–1925. [2] S. Darke, W. Hall, S. Kaye, J. Ross, J. Duflou, Hair morphine concentrations of fatal diacetylmorphine overdose cases and living diacetylmorphine users, Addiction 97 (2002) 977–984. [3] D.A. Engelhart, A.J. Jenkins, Detection of cocaine analytes and opiates in nails postmortem cases, J. Anal. Toxicol. 26 (2002) 489–492. [4] C. Cattaneo, F. Gigli, F. Lodi, M. Grandi, The detection of morphine and codeine in human teeth: an aid in the identification and study of human skeletal remains, J. Forensic Odontostomatol. 21 (2003) 1–5. [5] N. Raikos, H. Tsoukali, S.N. Njau, Determination of opiates in postmortem bone and bone marrow, Forensic Sci. Int. 123 (2001) 140–141. [6] F. Moriya, Y. Hashimoto, Distribution of free and conjugated morphine in body fluids and tissues in a fatal diacetylmorphine overdose: is conjugated morphine stable in postmortem specimens? J. Forensic Sci. 42 (1997) 736–740. [7] R. Pacifici, S. Di Carlo, A. Bacosi, S. Pichini, P. Zuccaro, Pharmacokinetics and cytokine production in diacetylmorphine and morphine-treated mice, Int. J. Immunopharmacol. 22 (2000) 603–614. [8] T.T. Noguchi, G.R. Nakamura, E.C. Griesemer, Drug analysis of skeletonizing remains, J. Forensic Sci. 23 (1978) 490–492. [9] C.L. Winek, E.M. Morris, W.W. Wahba, The use of bone marrow in the study of postmortem redistribution of nortriptyline, J. Anal. Toxicol. 17 (1993) 93–98. [10] T. Kojima, I. Okameto, T. Miyazaki, F. Chikasue, M. Yashiki, K. Nakamura, Detection of metamphetamine and amphetamine in a skeletonized body buried for 5 years, Forensic Sci. Int. 31 (1986) 93–102.

[11] K. Kudo, H. Sugie, N. Syoui, K. Kurihara, N. Jitsufuchi, T. Imamura, N. Ikeda, Detection of triazolam in skeletal remains buried for 4 years, Int. J. Legal Med. 110 (1997) 281–283. [12] H. Maeda, S. Oritani, K. Nagai, T. Tanaka, N. Tanaka, Detection of bromisovalum from the bone marrow of skeletonized human remains: a case report with comparison between gas chromatography/mass spectrometry (GC/MS) and high-performance liquid chromatography/mass spectrometry (LC/MS), Med. Sci. Law 37 (1997) 248–253. [13] C.L. Winek, A.G. Costantino, W.W. Wahba, W.D. Collom, Blood versus bone marrow pentobarbital concentrations, Forensic Sci. Int. 27 (1985) 15–24. [14] C.L. Winek, M. Pluskota, W.W. Wahba, Plasma versus bone marrow flurazepam concentration in rabbits, Forensic Sci. Int. 19 (1982) 151–163. [15] C.L. Winek, S.E. Westwood, W.W. Wahba, Plasma versus bone marrow desipramine: a comparative study, Forensic Sci. Int. 48 (1990) 49–57. [16] T. Nagata, K. Kimura, K. Hara, K. Kudo, Metamphetamine and amphetamine concentrations in post mortem rabbit tissues, Forensic Sci. Int. 48 (1990) 39–47. [17] I.M. McIntyre, C.V. King, M. Boratto, O.H. Drummer, Post-mortem drug analyses in bone and bone marrow, Ther. Drug Monit. 22 (2000) 79–83. [18] K. Terazawa, T. Takatori, Determination of aminopyrine and cyclobarbital from a skeleton by radioimmunoassay, J. Forensic Sci. 27 (1982) 844–847. [19] L.Y. Gorczynski, F.J. Melbye, Detection of benzodiazepines in different tissues, including bone, using a quantitative ELISA assay, J. Forensic Sci. 46 (2001) 916–918. [20] Y. Gaillard, G. Pepin, P. Marquet, P. Kintz, M. Deveaux, P. Mura, Identification et dosage de la benzoylecgonine, cocaı¨ne, me´thylecgonine-ester, code´ine, morphine et 6-ace´tylmorphine dans le sang total, Toxicorama VIII (1996) 17–22. [21] P. Kintz, M. Mangin, Simultaneous determination of opiates, cocaine and major metabolites of cocaine in human hair by gas chromatography/mass spectrometry (GC/MS), Forensic Sci. Int. 73 (1995) 93–100. [22] G.R. Nakamura, J.L. Thornton, T.T. Noguchi, Kinetics of diacetylmorphine deacetylation in aqueous alkaline solution and in human serum and whole blood, J. Chromatogr. 110 (1975) 81–89. [23] J.G. Umans, C.E. Inturrisi, Pharmacodynamics of subcutaneously administered diacetylmorphine, 6-acetylmorphine and morphine in mice, J. Pharmacol. Exp. Ther. 218 (1981) 409–415. [24] A.Y. Salmon, Z. Goren, Y. Avissar, H. Soreq, Human erythrocyte but not brain acetylcholinesterase hydrolyses heroin to morphine, Clin. Exp. Pharmacol. Physiol. 26 (1999) 596–600. [25] S. Schwarzinger, M. Hartmann, P. Kremminger, N. Muller, Hydrophobic forms of morphine-6-glucosides, Bioorg. Med. Chem. Lett. 11 (2001) 1455–1459.

Forensic Science International 166 (2007) 145–154 www.elsevier.com/locate/forsciint

Performance of the FearID earprint identification system Ivo Alberink *, Arnout Ruifrok Netherlands Forensic Institute, Department of Digital Technology, P.O. Box 24044, 2490 AA, Den Haag, The Netherlands Received 7 July 2005; accepted 3 May 2006 Available online 12 June 2006

Abstract The Forensic Ear Identification (FearID) research project was started in order to study the strength of evidence of earprints found on crime scenes. For this purpose, a sample of earprints from 1229 donors over three countries was collected. From each donor three left and three right earprints were gathered. On the one hand, operators denoted contours of the earprints to facilitate segmentation of the images, on the other anthropological specialists denoted anatomically specific locations. On the basis of this, methods for automated classification were developed and used for training of a system that classifies pairs of prints as ‘matching’ or ‘non-matching’. Comparing lab quality prints, the system has an equal error rate of 4%. Starting from a reference database containing two prints per ear, hitlist behaviour is such that in 90% of all query searches the best hit is in the top 0.1% of the list. The results become less favourable (equal error rate of 9%) for print/mark comparisons. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Earprint identification; Verification; Classification; Performance

1. Introduction In recent years, expert court testimony on earmarks found at crime scenes relied on the assessment of expert witnesses on biological ‘uniqueness’ of characteristics found on the marks, cf. [1]. Examples of these are overall shape and size, and details such as Darwinian tubercles, creases, moles, piercings or scars. Because of a relative lack of scientific basis, and the subjective nature of the assessments by the experts, reliability of earmark identification has been under fire. A good review article on the (lack of) scientific research up to 1999 with respect to earmark identification can be found in [2]. This has, e.g. resulted in rejection of earmark evidence in the State v. Kunze case in the United States, see [3], and the calling of a retrial in the Regina v. Mark Dallagher case in the United Kingdom, see [4] and [5]. To solidify the scientific basis for earprint/earmark identification, in the period 2002–2005 the EU financed Forensic Ear Identification (FearID) project was held, divided over nine institutes, including police academies, universities, the Netherlands Forensic Institute and two commercial partners, in Italy, the Netherlands and the UK. The project aimed at obtaining estimators for the strength of evidence of

* Corresponding author. Tel.: +31 70 888 6404; fax: +31 70 888 6559. E-mail address: [email protected] (I. Alberink). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.001

earmarks found on crime scenes and the development of methods to match and classify earprints. The ultimate goal of this is verification (1 to 1 matching) and individualisation (1 to many matching). Note that the goal was not to prove ‘uniqueness’ of earprints. The concept of uniqueness expresses the notion that given two earprints of high enough quality, a qualified enough expert will be able to decide with 100% certainty whether these originated from the same ear or not. From a statistical point of view, this notion is fundamentally unprovable and hence meaningless. What we can do is use the information present in earprints, and try to come up with a system that functions as well as possible in the matching of pairs of prints. Strength of evidence will ultimately be given by means of likelihood ratios. In the three countries, a training database was gathered of 1229 donors, donating three left and three right earprints each. Standard operating procedures were designed for the recovery and lifting of donor earprints, laid down in [6]. The document contains directions and instructions guiding the technician in collecting earprints. An important difference to earlier practice is that earprints were not gathered by making donors apply different amounts of force to special glass or flat surface. Starting from the notion that in practice a perpetrator will be listening for (the absence of) a sound, and hence uses a stable ‘functional force’, donors were instructed to listen for a sound supplied behind a glass plate. Earprints were then recovered

146

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

Fig. 1. Example: original print, clicked polyline (apex point to earlobe) and calculated ‘superstructure’. Table 1 Numbers of donors, collected prints and simulated marks

Donors Prints Marks

UK

NL

IT

Total

349 2089 30

430 2578 71

450 2697 115

1229 7364 216

UK, United Kingdom; NL, Netherlands; IT, Italy.

Fig. 2. Example: anatomically annotated print.

from the glass plate. For non-willing donors, this has the advantage that one may check whether ‘functional force’ was used by inquiring about the content of what was heard. For every 10th donor, either 1 or 2 simulated crime scene marks were taken. Next to this, an operational system was developed allowing for scanning, storing and processing of earprints. An example of this processing is the following. Users may add polylines to the digitized earprint images following the imprint of the ear. From these, connected structures are determined supposedly representing the imprints and which are referred to as superstructures (see Fig. 1). On the basis of these superstructures, further analysis is performed using image processing techniques. Another example of manual annotation is that anthropological experts added anatomical annotations of minutiae and landmarks present in prints as in Fig. 2. On the basis of this knowledge-based approach, comparisons of pairs of prints were then made. The anthropological workpackage of the FearID project decided on a set of possible minutiae and landmarks, laid down in [7], and the prints gathered in the database were annotated accordingly. In this way, an effort was made to

objectify (necessarily subjective) expert witness opinions, and lay a basis for classification of earprints on the basis of biological knowledge (as opposed to information acquired by (semi-)automatic image processing). In Section 2, the design of the main FearID sample is explained and concrete numbers are given. On the basis of the above, at the Netherlands Forensic Institute and University of Huddersfield three methods for extracting potentially useful features in the matching process were implemented. In Section 3, the ways in which performance is reported (equal error rates and hitlist behaviour of the system) and in Section 4 the features at the basis of the matching process are described. In Section 5, the analysis of the data is given, consisting of the training of the system and the validation process, looking at print/print and print/mark comparisons. In Section 6 the results are summarized. 2. Design of the database The database consists of 7364 prints of 1229 donors, taken according to the standard operating procedures. This means that all donors provided first three left, then three right earprints1 on the FearID ‘listening box’, which would be consecutively lifted using Black Gel Lifter. For every 10th donor, either 1 or 2 ‘uncontrolled’ prints would be taken, without instructions to the donor, to function as simulated crime scene marks, which are called ‘(simulated) marks’ from here on. In this way, over the three countries, the following numbers of prints were collected as described in Table 1. For a number of prints, the quality was so low that the necessary annotations could not be added. This happened for 1

In 10 cases 1 of the 6 prints per donor was missing.

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

147

Table 2 Numbers of prints and simulated marks removed from the sample for quality reasons

Prints Marks

UK removed

NL removed

IT removed

Total removed

103 (5%) 2 (7%)

143 (6%) 10 (14%)

137 (5%) 8 (7%)

383 (5.2%) 20 (9.3%)

UK, United Kingdom; NL, Netherlands; IT, Italy.

Table 3 Numbers of prints and simulated marks in the training and validation part of the database (before removal of poor quality prints and ‘marks’)

Donors Prints Marks

Training sample

Validation sample

771 4621 0

458 2743 216

example in case of empty or double prints. As in the real life situation, these prints were removed from the analysis. Of all prints, in this way a total of 383 prints, or 5.2% of the total, were removed, and 20 out of 216 (9.3%) simulated earmarks were taken out (Table 2). In the cases in which it was possible, prints were annotated as follows. First the operator gathering the prints would (digitally) add a polyline to the scan following the earprint contour. In second instance, an anthropologist from a different country would annotate the print anatomically. On the basis of the prints and consecutive manual annotations feature extraction and numerical analysis was performed. For training and testing of the model for classification, the database was divided into two parts, of 771 and 458 donors, respectively. The first part was used for training of the feature extraction methods and their numerical optimal combination. The second part was used for validation of the derived system. Training took place on pairs of prints gathered under lab conditions only, so all donors having simulated marks available were part of the validation part of the database. In Table 3, concrete numbers are given. 3. Performance measures With respect to evidential value there are two key concepts, that of verification, or 1 to 1 comparison, and that of individualization, or 1 to n comparison. The corresponding performance measures are equal error rates and hitlist behaviour, which we shall use further to do training and express performance of the system. 3.1. Equal error rates (EER) We start with the concept of verification, or 1 to 1 matching. A verification system is a classification system with two classes of outcomes: matching (or positive or acceptance) and nonmatching (or negative or rejection). Given the features in a system, for any pair comparison, a single value is constructed optimally summarizing the matching information. Classifica-

Fig. 3. Plot of the FRR vs. FAR, with the corresponding EER.

tion takes place according to whether the outcome does or does not exceed some threshold t. Common performance parameters with this type of system are the probabilities of making a wrong judgement, expressed in the false rejection rate (FRR) (cases in which the system declares a non-match in case of matching prints) and false acceptance rate (FAR) (cases in which the system declares a match in case of non-matching prints). Since the FRR and FAR are threshold-dependent, we rather use the equal error rate (EER), which is the (common) probability of misclassification starting from the threshold t for which FAR(t) = FRR(t). An example of this is given in Fig. 3. 3.2. Hitlist behaviour The second way in which the performance of the system is tested is by looking at hitlist behaviour of the system. Here, query prints are compared to a reference database and for any query print the relative position of the corresponding matching print that comes up highest in the hitlist is noted. This corresponds to the real life situation that one has a reference database of lab quality suspect prints, with, e.g. two prints per subject, and a crime scene mark is presented. This is also called 1 to n comparison. A possible result is depicted in Fig. 4. In the current setting, for the validation part of the sample, we have two reference databases, for left and right prints, respectively, that consist of two prints per ear, and use the remaining prints as query prints to test the system. We are not so much looking for a matching print of the query print at place 1 of the hitlist, but rather in, say, the top 20, to facilitate searching

Fig. 4. Example of hitlist following a query search, with two matching prints at places 4 and 15 and the best hit at place 4.

148

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

in a database. Notice that the fact that two prints per ear are included in the databases artificially enhances the success rate on ‘best’ hits. The latter is defendable in a real life database scenario though. We report hitlist results in terms of the worst behaviour of 90 and 95% of all the query searches, that is, for example, we may report that for 90% of all searches, the best hit is always in the top 2% of the hitlist. 4. Feature extraction In order to get to useful features for the matching process, three different methods were used. We describe, in turn, comparison of superstructures by means of weighted width and angular development, and comparison of anatomical annotation by the method of Vector template matching. 4.1. Weighted width comparison The first way of analysis is through comparison of superstructures. The method is related to that of ‘template matching using cross correlation’ which can be found in [8]. We start by giving a general explanation about the method. After this, we explain some details about the implementation (downsampling of the signals, maximum translation allowed for the cross correlation). 4.1.1. Method On the basis of the clicked polyline, representing a skeleton for the ear imprint, and the earprint scan, the area of the print

taken up by the imprint of the ear itself is reconstructed, as in Fig. 1. This area, in Fig. 1 on the right, is called ‘superstructure’ or ‘super-helix’ and on the basis of local intensities, a ‘medial axis’ is calculated for it. After this, the superstructure is ‘stretched’ along this axis, transforming the scan as shown in Fig. 5. In case of the same earprint, differently centered and rotated, identical stretched images will (ideally) appear. In this way, the stretching of superstructures solves the problem of translation and rotation (in)variance, which makes direct comparison of calculated superstructures feasible. Starting from the stretched superstructures, we look at the width of the superstructure along the medial axis, weighted by the corresponding intensities. In this way we compensate for the fact that at times width of the superstructure does not represent a part of the imprint of the ear at all, and is just there to keep up the continuity of the superstructure. The resulting signals are referred to as weighted width signals, see the third picture of Fig. 5. When considering two different prints of the same ear, it may occur that part of the ear makes an imprint on the one but not on the other print. In such cases, the stretched superstructure of the one print may be longer at one side. However, there will be an overlap between the two stretched structures, which can be matched optimally over each other. For this we only need to look for the optimal horizontal translation matching the two, which simplifies computational problems considerably. When comparing two signals, we look for the optimal translation to fit them on top of each other, optimizing correlation on the overlapping part. We use [1  optimal

Fig. 5. Image, masked by its superstructure, its stretched version and the corresponding width signal, weighted with respect to intensities.

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

149

Fig. 6. Weighted width signals of two prints from the same ear, matched optimally.

correlation] as the distance between two prints. This equals 0 for perfectly correlated signals and 2 for non-correlated ones. An example of an optimal fit is given in Fig. 6. 4.1.2. Training Training of the method consisted of optimization of two parameters, namely that of a downsampling factor d on the signals and on the maximally allowed translation t of the one signal over the other. Downsampling of the signals takes place because it reduces complexity of the calculations, the resolution of the images (600 dpi, images of 2100  3000 pixels) being too high for the relatively coarse method. To optimize d and t the following was done. From the training part of the main sample, we took 10 random samples of 1000 matching and 1000 non-matching prints. On the basis of particular choices for the two parameters, the equal error rates (see Section 3.1) involved were determined, and used as criterion for optimization. Details may be found in [9]. We find that downsampling of the signals by a factor 20 does not lead to a decrease in EER performance of the method. Staying on the conservative side, we furthermore used the combination d = 10 and t = 20.

4.2. Angular comparison 4.2.1. Method The second method to compare clicked prints consists of the comparison of signals keeping track of the angle of the medial axes corresponding to the superstructures with the x-axis, see for example the concept of ‘turning functions’ in [10]. We refer to the method as angular comparison. The resulting signals are fitted on each other by optimal translation. Comparison using this method is translation and rotation invariant. For each earprint, a segmented superstructure and medial axis have been determined. Starting at the anthelix branching and ending at the ear lobe, the angle this medial axis is making with the x-axis is stored as in Fig. 7. Given any angular signal, a vertical translation of the signal corresponds with a rotation with the same angle of the medial axis. When comparing two angular signals, horizontal translation of the one over the other corresponds to doing a match of the two using different starting points. The optimal fit of the resulting signals is determined by finding the horizontal and vertical translation minimizing the distance between the two. As distance measure the mean quadratic distance between the signals on their overlap is used, given the translation.

Fig. 7. Example: medial axis and its angular signal.

150

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

Fig. 8. On the left (A): illustration of the medial axis (in red) of an earprint superstructure, or super-helix. On the right (B): illustration of anthelix points. For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.

4.2.2. Training On the basis of the medial axes, angular signals are determined as follows. For every point p in the axis, we look at the angle made between the point c positions before p, p itself, and the point c positions behind p, for some constant c. This forms the basis for the increase of the angular signal. When comparing angular signals, we look for an optimal translation of the one signal on the other, minimizing the absolute distance between the two on the overlapping part. (Optimizing cross correlation of signals as for the weighted widths leads to a variable which has much less discriminative power.) There is a bound on this translation in the sense that starting from the situation where the apex points coincide, the signal may be translated maximally t positions. A third parameter is yielded by downsampling the medial axes by a factor d before doing any of the above. Details of the training may again be found in [9]. Staying on the conservative side, we chose a downsample factor of 20, so that approximate lengths of the signals are 200 rather than 4000, which has no negative effect on EERs. We used the combination d = 20, c = 10 and t = 20. 4.3. Vector template matching The third method used to do print comparisons was based on anatomical annotation of earprints, analysed through the method Vector template matching (VTM). Following its anatomical annotation, each print has a template consisting of labelled points representing earprint landmarks and minutiae, distinguished into different classes. Prints are compared by assessment of the similarity between their templates. We describe the annotation process and the concept of Vector template matching. 4.3.1. The annotation process Subsequent to print collection, earprints were annotated in a three-stage process for which FearID’s Earprint Storage and Analysis System was used. The first stage was carried out at the collection sites Centrex NTC (UK), LSOP (Netherlands) and Padova University (Italy). Stages two and three were carried out independently of the collection site, in a different country from the site where they were collected. All annotations were based on a standard set and agreed by the anthropological analysis

group that consisted of Glasgow University, Leiden University Medical Centre and Padova University. The first annotation stage consisted of operators clicking the initial axis of the superstructure or super-helix structure used to represent the visible features of the ear palm (or pinna) (see Figs. 1 and 8A). The superstructure is a spiral-like structure, starting at the anthelix or anti-helix, and ending at the lobule, chosen as appropriate for subsequent computerised analysis of print patterns. The initial axis is a piecewise linear path linking points placed along the superstructure, with additional specific points representing the lowest point of the lobe and, in the anthelix area, apex and indentation markers, indicating the number and extent of anthelix branches (see Fig. 8B). During the second stage of print annotation, operators added print transition lines. These indicate the transition between print segments one encounters when following the superstructure path and contain information about superstructure segmentation. In the third and final stage the anthropological analysis group annotated anatomical features, minutiae, landmarks and some other characteristics. Here minutiae are characteristic anatomical details such as moles, Darwinian nodules, piercings, or particular crease formations at particular positions, and the term landmark is used for particular predefined points on gross features or on the outline of lacunae, unlikely to change through time. Subdivision along these lines and inclusion of features from the first and second annotation stage led to a number of 104 features used in the VTM analysis. A full description of the instructions for marking minutiae and related landmarks in earprints is given in [7]. 4.3.2. Vector template matching analysis Depending on the nature of a particular earprint feature, operators used either a point, line or area marker to highlight landmarks and minutiae. Point minutiae were included in the point matching set of a print. Line and area landmarks were processed to extract salient points such as extremities and bifurcation points of creases and barycentres of papules, thus reducing annotated line and area features to characteristic points as well. The coordinates were stored for each print, including the feature names (label), and formed the earprint

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

Fig. 9. Example of two labelled point patterns, differing only in translation (268).

point pattern used for matching earprints. The patterns are represented as lists of the form: P  ½ p1 ; p2 ; . . . ; pi ; . . . ; pNP ; ðPÞ

ðPÞ

ðPÞ

where list item pi ¼ hxi ; yi ; ‘i i gives the Cartesian coordinates of the ith annotated point in the digitized image, together with the label indicating the nature of the point. The ordering of items in the points list is arbitrary. The coordinates of corresponding points in different prints from the same ear will naturally differ because of having different origin and rotation, so the comparison mechanism needs to be invariant to translation and rotation. Our numerical analysis is based on a method referred to as Vector template matching, described in detail in [11]. As described, each print has a template consisting of labelled points representing annotated earprint landmarks and minutiae, distinguished into different classes. Prints are compared by assessment of the similarity between their templates. In the case of prints originating from the same ear, the same labels are expected to turn up, though because of translation and rotation of the ear not on the same coordinates. Comparison of the templates takes place in the following way. For any vector in print 1, all vectors in print 2 are determined sharing the same (anatomically meaningful) labels. In Fig. 9 this would for example mean starting at the vector with the labels A and B in print 1, and comparing this to the corresponding vector found below in print 2.

151

The angle between the vectors, here approximately 268, is then determined and stored. In order to minimize comparison of vectors that do have matching labels but do not correspond, the ratio of the length of the vectors – which anatomically is supposed to be close to 1 – is supposed to be inside the fixed interval (0.90;1.11), which is the result of independent training of the method. The above is done for all combinations of vectors from print 1 and print 2 sharing the same labels and approximate length, and a histogram is made of the observed outcomes. In the case of two matching prints, for the histogram this will result in a peak near the actual rotation of the one print with respect to the other, and low variation in outcomes. In the case of non-matching prints, the histogram is expected to be noisy (see Fig. 10). Two point patterns are shown to be similar by assessing the dominant mode of the distribution of the angles between pairs of vectors from the two patterns. To this, for any VTM comparison, nine features were extracted from the resulting histogram: 1. 2. 3. 4. 5. 6. 7. 8.

N, total number of comparisons of vectors in the procedure. Std, standard deviation of resulting histogram. IQR, inter quartile range of resulting histogram. Peak, peak value of resulting histogram. Peak/N. Peak/Std. Peak/IQR. Ntot, total number of possible comparisons of vectors in the VTM procedure, not considering either labels or vector length ratio. 9. N/Ntot. For details about the implementation and training (with respect to histogram bin width and vector length ratio) of the VTM method, see [11]. 5. Analysis The analysis of the outcomes is based on the statistical method of (binary) logistic regression and is described in

Fig. 10. Example of histograms corresponding to a pair of matching prints (on the left) and a pair of non-matching prints. In the case of matching prints we see a compact histogram, in this case around rotation angle 08. In the case of non-matching prints the histogram is much more ‘noisy’.

152

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

Section 5.1. Performance of the system on the validation part of the Main sample is given in Section 5.2. 5.1. Selection of discriminant score (based on logistic regression) We aim to obtain a score function that minimizes the number of features used, but is still close to optimizing the EER. The reason behind this is that especially the nine VTM parameters are all highly correlated, and there is a risk of overtraining of the system on the data. The analysis of the outcomes is based on the statistical method of (binary) logistic regression or BLR, cf. [12]. Binary logistic regression is used rather than linear discriminant analysis since normality assumptions on the features are violated and some features are discrete, in which case BLR outperforms linear discriminant analysis. Based on the training data, the BLR method extracts a linear combination of certain of the used features, optimally separating pairs of matching from pairs of non-matching prints. We denote the distances obtained by the weighted width and angular method by ww and ang, respectively. For reasons of symmetry, we added the inverse values and natural logarithms (from here on: ln) to the mentioned features, thus ending up with 33 features per comparison of prints. However, no more than one instance of the same feature was used in the end model for the score function: for example, either N, 1/N or ln(N). Comparisons for which the VTM algorithm led to N < 2 were filtered out since in that case the histogram lacked informative value. Note that in practice this is what one would do as well. For the training sample, there were a number of circa 5.3 million possible comparisons of non-matching prints and 3084 of matching prints, on the basis of which the BLR score was trained. Because of filtering of bad, double and empty prints, as well as comparisons with VTM results for which N < 2, the number of valid comparisons further decreased to 2727. Since BLR is not robust against differences in size between groups that are to be separated, 10 (partially overlapping) subsamples of 1000 matching and 1000 non-matching pairs of prints were taken, sampled independently. The BLR analysis was performed such that for each of these subsamples, first the system of 33 features was trained, using low thresholds for removal of features. For this, we used the SPSS module for BLR, backward: LR method, with the parameters settings ‘Probability for stepwise Entry and Removal’ both at 0.01. The outcomes for the 10 different training samples were compared, and it turned out that ang1, ln(ww), Peak and ln(Ntot) together lead to a steady EER (over all 10 training samples) of 4.8%. It was decided to further use the discriminant score

Table 4 Numbers in the validation sample Total number of

Left

Right

Query prints Reference prints (Simulated) marks

424 900 96

412 886 100

5.2.1. Numbers The validation sample consists of 457 donors with 3 prints per ear. Prints of bad quality, empty prints, or prints lacking the necessary annotations were removed, leading to a total of 2639 ear prints (out of a possible 2742). For 119 of the donors moreover, either 1 or 2 (simulated) earmarks were taken, in order to simulate print/mark comparisons, as opposed to comparing prints that are both of lab quality. The objective is to resemble the real life situation better, and 199 of these simulated crime scene marks were taken. To facilitate investigation of hitlist behaviour, the remaining prints were divided into reference databases and collections of query prints as follows. For left and right ears, separate databases were built. For a fixed ear side, for any ear from the sample, in case of three remaining prints, two of these were included in the reference database and the remaining one in the collection of query prints. In case of two remaining prints, both were included in the reference database, and in case of less than two, the ear was disregarded completely. In this way, for both left and right ears, a reference database was constructed consisting of pairs of two prints from all applicable ears, and a collection of query prints having no overlap with the reference database, which however always contains matching prints. Numbers are given in Table 4. 5.2.2. Performance of the system 5.2.2.1. EER. We compared all (lab) query prints to all reference prints (per ear side). Here we find an EER of 3.9%. In more detail this is illustrated by the histograms in Fig. 11.

D ¼ 1=1 þ expðLÞ; where L = 20 + 0.4 ang1  3.4 ln(ww) + 0.24 Peak  3.4 ln(Ntot). 5.2. Validation of the model After determination of the discriminant score, the validation print set of the main database is used to test its performance.

Fig. 11. Histograms of discriminant scores for matching (above) and nonmatching pairs of prints (below).

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

153

Table 7 Hitlist behaviour of the system Position

Fraction of searches (%)

1 2 3 4 5 6 7 8 9 10 >10

Fig. 12. Histograms of discriminant scores for matching (above) and nonmatching pairs (below) of prints vs. marks.

Comparison of the simulated marks with the reference databases leads to an EER of 9.3%. The histograms are shown in Fig. 12. Figs. 11 and 12 show that for non-matching prints, almost the same probability distribution is underlying the whole, and that for matching prints, the behaviour of print–mark comparisons is much more erratic. Combined EER results, also per ear side, are summarized in Table 5. Note that the EERs are significantly larger in the print/mark comparison scenario than for comparison of lab prints, and that discrimination of right ears works better than that of left ones. To give an idea about performance of the system on the basis of only weighted width and angular comparison, not using anatomical annotations, we give the EER results using the discriminant score D = 1/(1 + exp((7.2 + 0.68 ang1  3.6ln(ww)))) in Table 6. 5.2.2.2. Hitlists. The second way in which the performance of the system is tested is, as explained in Section 3.2, by constructing hitlists for every query print compared to its reference database. For any query print, we look for the relative position of the corresponding matching print that comes up highest in the hitlist. For a left query print this means it is Table 5 EER performance of the system Side of ear

Print–print (%)

Print–mark (%)

Overall Left Right

3.9 4.7 3.0

9.3 11.4 7.0

Table 6 EER performance of the system solely based on image processing, so not using anatomical annotations

EER

Print–print (%)

Print–mark (%)

8.5

15.2

Print–print

Print–mark

91 3.3 1.0 0.6 0.5 0.6 0 0 0.1 0 2.2

72 8.6 3.2 1.0 1.6 2.2 1.1 4.3 0 0 5.9

Table 8 Hitlist behaviour of the system as a function of side and print type Side

Print–print (%)

Print–mark (%)

90% of best hits in top Overall Left Right

0.1 0.1 0.1

2.0 6.5 2.1

95% of best hits in top Overall Left Right

0.2 0.2 0.2

7.3 7.3 3.5

compared to all 900 prints in the left earprint database, and the best position of the matching prints in the resulting hitlist is taken. Overall results are presented in Table 7. Comparing prints to prints, here we see for 90% of the query prints that the best resulting hit will be in the top 0.1% (in the case of 900 left prints: number 1) of the hitlist. Results split out in left and right can be found in Table 8. Again the phenomenon occurs of right earprints performing better than left ones, which is probably because they are of better quality. We conjecture this is correlated to left or righthandedness of the donors (who will mostly be right-handed). 6. Summary In recent years, forensic individualization based on earmarks has been under fire. To solidify the scientific basis for earprint/ earmark identification, the EU financed Forensic Ear Identification project was started in nine institutes over Italy, the UK and the Netherlands. A training database was gathered of 1229 donors each donating three left and three right prints. Furthermore, methods for automated classification on the basis of manual annotations were developed. The manual annotations were two-fold: on the one hand, of operators denoting the contour of the earprints to facilitate segmentation of the image, on the other, of anthropological specialists denoting specific anatomical locations in earprints (and marks). On the basis of this, a system was trained to classify pairs of

154

I. Alberink, A. Ruifrok / Forensic Science International 166 (2007) 145–154

prints as matching or non-matching. Comparing lab quality prints to one another, the resulting matching system has an equal error rate of 4%, and starting from databases containing two prints per ear, hitlist behaviour is such that in 90% of all query searches the best found hit is in the top 0.1% of the list. The above is rather satisfying for a semi-automated system. However, one has to take notice of the following facts. On the one hand there is the issue of manual annotation. In the system, for any one of the two annotation procedures, the same operator annotated all prints of any one donor. That is to say, given donor A, his or her prints were annotated with respect to contour of the imprint by fixed operator B, and anatomically annotated by fixed operator C. Because of this, matching prints can appear more alike in the training system than they are in reality. To investigate the latter, two inter-operator experiments were performed: cf. [13] and [14]. Besides this, there is the fact that comparison of lab quality prints is bound to be more stable than comparison of suspect prints to crime scene marks. To this, gathering of simulated earmarks took place for every 10th donor and performance of the system for print/simulated earmark comparisons was checked as well. According to experts, simulated earmarks were comparable in quality, or worse, than real-life crime scene marks, and in this respect the drop of equal error rate from 4 to 9% seems to be informative. From a technical point of view we note that our image procesing techniques work just as well on images with dimensions 210  300 as the original 2100  3000. We expect that the system may be improved by means of more optimal feature extraction (image processing). Acknowledgements This study was carried out within the framework of the FearID project, which is a shared-cost RTD project funded under the fifth Framework Programme of the European Community, within the Competitive and Sustainable Growth Programme, Measurement and Testing Activity, Contract G6RD-CT-2001-00618. Authors would like to thank Ruud

van Basten (LSOP, the Netherlands), Francesca De Conti, Marta Giacon (University of Padova, Italy), Zale Johnson (Central Police Training and Development Authority, Durham, UK), Lynn Meijerman (University of Leiden, the Netherlands), and Sarah Sholl (University of Glasgow, UK) for annotation of the sample. Moreover, they want to thank Hartmut Kieckhoefer (University of Huddersfield, UK) for implementation of the VTM method and help on that of weighted width comparisons. References [1] C. van der Lugt, Earprint Identification, Elsevier Bedrijfsinformatie, Den Haag, 2001. [2] C. Champod, I.W. Evett, B. Kuchler, Earmarks as evidence: a critical review, J. For. Sci. 46 (2001) 1275–1284. [3] S.v. Kunze, Court of Appeals Washington, Division 2, 97 Wash. App. 832, 988 P.2d, 977, 1999. [4] Ear Print Catches Murderer, http://news.bbc.co.uk/1/hi/uk/235721.stm, BBC online network, 15-12-1998. [5] Man Convicted of Murder by Earprint is Freed, http://www.timesonline.co.uk/article/0,1-973291_1,00.html, TIMES ONLINE, 22.01.2004. [6] Z. Johnson, Standard Operating Procedure for the Taking of Earprints. Internal FearID Report, 2003, http://forensic.to/fearid/Procedure.doc. [7] F. De Conti, M. Giacon, L. Meijerman, S. Sholl, Instructions for Marking Minutiae and Related Landmarks in Earprints. Internal FearID Report, 2004, http://forensic.to/fearid/Labelling_minutiae_and_characteristic_landmarks.pdf. [8] John C. Russ, The Image Processing Handbook, second ed., CRC Press, 1995. [9] I.B. Alberink, A.C.C. Ruifrok, Operator effects in the acquisition of earprints, personal communication. [10] E.M. Arkin, L.P. Chew, et al., An efficiently computable metric for comparing polygonal shapes, TPAMI 13 (3) (1991) 209–216. [11] H. Kieckhoefer, M. Ingleby, I.B. Alberink, Vector Template Matching of Earprints Using Labelled Point Features. Internal FearID Report, 2005, http://forensic.to/fearid/VTMfinal.doc. [12] F.C. Pampel, Logistic regression: a primer. Sage university pagers series on quantitative applications in the social sciences, series no. 07-132, 2000. [13] I.B. Alberink, A.C.C. Ruifrok, Inter-operator test for the clicking of polylines in earprints, personal communication. [14] I.B. Alberink, A.C.C. Ruifrok, H. Kieckhoefer, Inter-operator test for anatomical annotation of earprints, J. For. Sci., 2006.

Forensic Science International 166 (2007) 155–163 www.elsevier.com/locate/forsciint

Male amelogenin dropouts: phylogenetic context, origins and implications A.M. Cadenas a,1, M. Regueiro a,1, T. Gayden a, N. Singh b, L.A. Zhivotovsky c, P.A. Underhill d, R.J. Herrera a,* a

Department of Biological Sciences, Florida International University, University Park, OE 304, Miami, FL 33199, USA b Faculty of Medicine, Henan University, Kaifeng, Henan, China c N.I. Vavilov Institute of General Genetics, Russian Academy of Sciences, Moscow, Russia d Department of Genetics, Stanford University, Stanford, USA Received 1 March 2006; received in revised form 3 April 2006; accepted 7 May 2006 Available online 15 June 2006

Abstract Several commercial PCR multiplex kits incorporate the amelogenin locus for the purpose of human gender identification. Consequently, erroneous results in the electropherogram profile of this locus can carry important forensic implications. In this study, dropout of the amelogenin Y allele was detected in 5 out of 77 phenotypically normal Kathmandu males using the AmpFlSTR1 Identifiler1 kit. A battery of male-specific markers including SNPs, STRs, STSs, and a minisatellite were amplified for the five amelogenin null samples in order to delineate the breakpoints of the deletions as well as assess the overall integrity of the Y-chromosome. This study represents the first to examine the haplogroup affiliation of the AMGY deletions. The analyses performed suggest a single origin for the five deletions as indicated by their allocation to a specific Yhaplogroup (J2b2-M241), related Y-STR haplotypes and identical regional localization of breakpoints. The age estimated from the microsatellite variation for the amelogenin deletions (if they are associated by descent) is approximately 6.5  3.3 ky, younger than the previously reported related age of the M241 haplogroup representatives (13–14 ky). Our data in combination with previous publications suggest a concentration of afflicted individuals in the Indian subcontinent, possibly as a result of common ancestry. The elevated incidence of the amelogenin dropout in these populations accentuates the need to utilize other loci for gender determination in order to obtain an accurate set of inclusion criteria in forensic casework. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Amelogenin; Deletion; Y-SNP; Y-STR; Forensic

1. Introduction Various commercial PCR multiplex reaction kits co-amplify the amelogenin locus in conjunction with autosomal STR loci for gender and individual identification, respectively. The amelogenin locus is a single copy gene with homologs present in the X- and Y-chromosomes (AMGX and AMGY, respectively) differing in both size and sequence [1,2]. Consequently this marker can be utilized for sex identification and has served to differentiate male from female genotypes in forensic casework and prenatal diagnosis. Two primer sets

* Corresponding author. Tel.: +1 305 348 1258; fax: +1 305 348 1259. E-mail address: [email protected] (R.J. Herrera). 1 These authors contributed equally to this article. 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.002

routinely employed to discriminate the AMGX and AMGY alleles take advantage of a 6 base pair (bp) deletion on the Xchromosome, generating X/Y amplification fragments of 106/ 112 bp or 212/218 bp [3–6]. Large-scale deletions/rearrangements have been reported within the Y-chromosome, especially in studies of the AZF euchromatin region [7,8], some of which have distinctive haplogroup assignment [9]. Similarly, anomalous results in the amelogenin PCR profile of individuals have also been reported, including cases of mutations in the primer-annealing region [10–12] and complete deletion of AMGY in males [13–18] but without phylogenetic context. The consequences of an incorrect gender determination, particularly in forensic applications where sex identification is crucial, are significant [19]. Additional implications exist for the medical field as well since adjacent genes may comprise the deleted region such as

156

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

the open reading frame for gonadoblastoma (GBY), postulated to be in the pericentric region between DYS267 and DYS270 [20], and azoospermia which occupies an extensive region downstream from the AMGY sequence associated with male infertility [21]. In the present study, dropouts of the amelogenin Yallele in 5 out of 77 males (6.49%), from Kathmandu, Nepal, are reported and characterized as deletions rather than point mutations in the primer-annealing region. The breakpoints were characterized at high resolution through the amplification of a battery of mapped Y-chromosome markers. In order to assess the ancestry and age of the deletions, a combination of Y-specific binary marker and Y-STR analyses were conducted. Although previous studies have provided Y-STR haplotypes for the amelogenin nulls encountered, they were primarily focused on

describing the existence and breadth of the deletions [15,16,18]. Our results restrict the AMGY deletions to a specific haplogroup and associate them regionally, consistent with the observation that Y-chromosome phylogenetic structure is often highly correlated with geography. In combination with the observed Y-STR haplotypes, an estimation of the approximate age of the deletions was also generated. Consequently, the comprehensive analysis performed provides information on various facets of the amelogenin deletions including high-resolution localization of the proximal and distal breakpoints relative to the centromere on the short arm of the Y-chromosome, integrity outside the deleted region and evolutionary history in terms of possible common ancestry as well as regional incidence in a global context.

Fig. 1. (A) Ideogram of the Y-chromosome. Cytogenetic designations for the bands are on the left. Lines to the right indicate positions of Y-binary markers and microsatellites analyzed. The boxed-in region spanning Yp11.2 to Yp11.1 is amplified for further detail in Panel B. (B) Position of AMGY with respect to other amplified loci: sY1091, Rep, sY1242, sY605, sY74, M12, MSY1, DYS257b, sY2180, SHGC-80640, sY56, sY1293, sY1078, and DYS19. The observed deletion encompasses sY1242, sY605, sY74, Amel, M12, MSY1, DYS257b and sY2180. The breakpoints of the deletion fall in the regions between sY1242 to Rep and sY2180 to SHGC-80640.

Table 1 Chromosomal location, amplification conditions and pertinent references for Y-specific STSs and SNPs examined Marker M410

Amplicon size (bp)

2.7

Reverse primer

Cycling conditions

Genbank Accession/ refSNP (rs#)

Reference

40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 1 94 8C, 2 min; 30 (94 8C, 1 min; 60 8C, 1 min; 72 8C, 1 min); 1 72 8C, 7 min 39 (94 8C, 30 s; 60 8C, 30 s; 72 8C, 1 min); 1 72 8C, 10 min 39 (94 8C, 30 s; 61 8C, 30 s; 72 8C, 1 min); 1 72 8C, 10 min 39 (94 8C, 30 s; 60 8C, 30 s; 72 8C, 1 min); 1 72 8C, 10 min 39 (94 8C, 30 s; 60 8C, 30 s; 72 8C, 1 min); 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 1 95 8C, 11 min; 28 (94 8C, 0.5 min; 65 8C, 1 min; 68 8C, 3 min), 1 68 8C, 30 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 35 (94 8C, 4.5 min; 70 8C, 2 min; 72 8C, 2 min), 1 72 8C, 7 min 39 (94 8C, 30 s; 60 8C, 30 s; 72 8C, 1 min) 1 72 8C, 10 min 39 (94 8C, 30 s; 56 8C, 30 s; 72 8C, 1 min) 1 72 8C, 10 min 39 (94 8C, 30 s; 61 8C, 30 s; 72 8C, 1 min) 1 72 8C, 10 min 39 (94 8C, 30 s; 60 8C, 30 s; 72 8C, 1 min); 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min 40 (94 8C, 1 min; 54 8C, 1.5 min; 72 8C, 4 min), 1 72 8C, 10 min

n/a

Sengupta et al. [23]

G66122

Lattanzi et al. [18]

n/a

Present study

G75488

Kuroda-Kawaguchi [29]

G65838

Tilford et al. [24]

G31361

Vollrath et al. [25]

G42828

Underhill et al. [26]

n/a

Chang et al. [16]

G38358

Vollrath et al. [25]

G66265

Tilford et al. [24]

AC017019.3

Sulston and Waterston [27]

G11989

Hudson [28]

G75511

Kuroda-Kawaguchi [29]

G66114

Tilford et al. [24]

G42821

Seielstad et al. [30]

rs8179022

Cinniog˘lu et al. [31]

NT_011875

Hammer and Horai [32]

G42825

Underhill et al. [26]

rs13447352

Cinniog˘lu et al. [31]

1161

6.432

Rep

288

6.437

gaatggtaccagctcctcc

actggatacctttctagga agaattg aacattggcccagaagc caggg caatgatcgccattctaactg

sY1242

440

6.443

cgtcggtattttacgacacg

gcatttgtttttcatgtgcg

sY605

279

6.65

acctccgaagactgaaccag

cccttgagtccacagagtcc

sY74

132–133

6.67

ttttagagtcattggccagg

M12

403

7.6

ctctgaaaaaaaggca gcag aaaagaagttgaggact ggagc attaattgaggttgttgtgc atacagat tgatacacttcctccttt agtgg ggttagggaaagtgtgcagc

288

8.6

actaaaacaccattagaaa caaagg acagaggtagatgctgaa gcggtatagc gaacttgtcgggaggcaat

86

8.75

ctccctgcaccatgactaca

SHGC-80640

306

8.93

tcatacccagagccaaatgagat

sY56

251

8.94

gactgccagcctcataaaaa

cgaaaaatgtgggagaa caagac tccagaaggcatgttaggaa

sY1293

834

9.2

tcccctcagccatctgtatt

ggcccacctgagtaatggta

sY1078

387

9.9

ataaatggtgcaaaccgcag

tggcatcaaagttctgcttg

DYS271

209

12.5

aggcactggtcagaatgaag

aatggaaaatacagctcccc

M241

366

13.4

aactcttgataaaccgtgctg

tccaatctcaattcatgcctc

YAP

150

20

caggggaagataaagaaata

actgctaaaaggggatggat

M9

340

20.1

gcagcatataaaactttcagg

aaaacctaactttgctcaagc

M304

527

21

ctagagggtattggggtaggc

ggaaataccttcacctaaata

MSY1 DYS257b sY2180

1500–2500

7.92

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

Forward primer caatcattgaccttaagtct gagtccc tctgcatgttctcctcaggt

sY1091

395

Position (Mb)

157

158

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

2. Materials and methods 2.1. Collection of samples and extraction of DNA Blood samples from 77 healthy, unrelated human males from Kathmandu, Nepal were collected in EDTA Vacutainer tubes. DNA was extracted from the blood using the phenol–chloroform extraction method [22]. The collection process was performed in strict compliance with IRB and NIH guidelines as well as with any additional regulations imposed by the institutions involved. When not in use, samples were stored at 80 8C 2.2. AmpFlSTR1 Identifiler1 analysis DNA amplification of the 15 STR loci and the amelogenin locus was performed using the AmpFlSTR1 Identifiler1 PCR Amplification Kit (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions with the recommended amount of DNA (0.5–1.25 ng) in a GeneAmp1 PCR System 9600 thermal cycler (Applied Biosystems, Foster City, CA). DNA fragment separation and detection was achieved in an ABI Prism 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA). As an internal size standard, ABI Genescan500 LIZ was utilized. Amplicon sizes were determined utilizing the GeneScan1 3.7 software and alleles were designated by comparison to an allelic ladder provided by the manufacturer using Genotyper1 3.7 NT software. 2.3. Amplification of AMEL-2F/3R In order to determine whether the anomalous PCR profile observed was caused by primer mismatch due to a point mutation affecting the oligonucleotide binding site(s) targeted by the amelogenin AmpFlSTR1 Identifiler1 primers, the amelogenin null samples were amplified with a different amelogenin primer pair (AMEL-2F/3R) following the conditions reported by Chang et al. [16]. The amplicons generated with the AMEL-2F/3R oligonucleotide pair encompasses the full region of the amelogenin primer-annealing sites employed by the Identifiler1 kit (see illustration in Fig. 1 of Roffey et al. [10]). The amplified products were separated by 1X TAE, 2% agarose gel electrophoresis and stained with ethidium bromide. The presence of two bands at 1174 and 985 bp denotes a male profile while a female profile displays a single band at 1174 bp. Designation of the X and Y alleles was determined by visual comparison with a 1 kb DNA ladder (NEB), a positive male control, and a female control. DNA profiles were visualized under UV light in a Fotodyne FOTO/Analyst1. 2.4. Amplification of Y-specific markers The deletions encountered in the amelogenin null samples encompass a segment on the short arm (p-arm) of the Y-chromosome. To further delineate its extent, the samples were challenged at markers distributed throughout this chromosome. The corresponding amplicon sizes, primer sequences, cycling conditions, Genbank accession numbers and references for these loci

are listed in Table 1. The amplified MSY1 and sY1091 products were separated on 1X TAE, 1% agarose gels, while 2% agarose gels were utilized for the remaining markers. The gels were subsequently stained with ethidium bromide and visualized by UV light in a Fotodyne FOTO/Analyst1 system. Reactions were carried out in the presence of positive male and female controls. 2.5. Haplogroup assignment In order to investigate the Y-haplogroup ancestry of the individuals exhibiting the deleted amelogenin sequence, six biallelic Y-markers (YAP, M9, M12, M241, M304, and M410) were genotyped following the most recent Y-tree topology [23] using a hierarchical system by standard methods, including PCR/RFLP and the YAP polymorphic Alu insertion (PAI) [32]. DNA fragments were separated by 1X TAE, 3% agarose gel electrophoresis followed by ethidium bromide staining and UV light photography. Results are given according to the Y-SNP haplogroup nomenclature suggested by the Y Chromosome Consortium [33]. 2.6. Microsatellite analysis To further investigate the ancestry of these deletions and estimate the age of the five mutations, 10 Y-specific microsatellite loci (specifically DYS19, DYS388, DYS389I, DYS389II, DYS390, DYS391, DYS392, DYS393, DYS439, and DYSA7.2) were typed as described by Cinniog˘lu et al. [31]. DYSA7.2 is also known as DYS461 [35]. The methodology for analysis of microsatellite variation within a binary haplogroup assumes a stepwise mutation model, whose value is based upon a generation time of 25 years [34]. The age of microsatellite variation was estimated as the average squared difference in the number of repeats between all current chromosomes and the founder haplotype, averaged over microsatellite loci and divided by w ¼ ð6:9  1:3Þ  104 (or approximately (2.8  0.5)  105 per locus per year) with the standard errors computed over loci [35,36]. Here, w is the effective mutation rate. 3. Results 3.1. Detection of AMGY dropouts Five phenotypically normal male individuals, were incorrectly genotyped as females using the AmpFlSTR1 Identifiler1 PCR Amplification Kit (Applied Biosystems, Foster City, CA), as demonstrated by the absence of the male-specific 112 bp peak and the singular presence of the female-specific 106 bp peak. Amplification of these samples with the AMEL2F/3R primers which flank the Identifiler1 amelogenin oligonucleotides did not produce AMGY PCR fragments, although the X allele was detected in the gel (Fig. 2). 3.2. Characterization of the deletions In order to delineate the location of the proximal and distal breakpoints (in respect to the centromere) of the deletions

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

159

Fig. 2. (A) Electropherograms of AmpFlSTR1 Identifiler1 results for an amelogenin null individual (Kat28), a female control and a positive male control. A dropout of AMGY is observed in the top panel generating the same electropherogram profile as the female control while the X and Y alleles are both present at the amelogenin locus for the positive male control. (B) PCR products using the AMEL-2F/3R primers. Lanes 1–5 contain the five male AMGY null individuals (Kat16, Kat28, Kat50, Kat62, and Kat66, respectively) and show a deletion of the 985 bp product. The positive male control (Lane M) displays the 1174 bp X allele as well as the 985 bp Y allele, while the female control (Lane F) has only the X product. Lane L contains the 1 kb DNA ladder and Lane N contains the negative control.

Fig. 3. Gel electrophoresis of sY56 and sY605. Lanes 1–5 correspond to the five AMGY deleted samples (Kat16, Kat28, Kat50, Kat62, and Kat66, respectively), Lane L to the ladder, lane F to the female controls, Lane M1 and M2 to the positive male controls, and Lane N to the negative controls. For sY56 the expected 251 bp PCR product is detected in the five amelogenin deleted individuals as well as the positive male controls, while the female control generates no PCR fragment. In contrast, the sY605 locus does not display any bands in the five samples or the female control despite obtaining the 279 bp product in the two positive male controls.

160

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

detected in the five Kathmandu males, a battery of Y-specific markers listed in Table 1 were examined. Fig. 1 illustrates their positions on the human Y-chromosome. Comprising a region of approximately 2.3 Mb, seven markers (sY1242, sY605, sY74, M12, MSY1, DYS257b and sY2180) did not yield PCR products in the deleted individuals despite observing the expected amplicons in the positive male controls. However, at just 5 kb from STS sY1242 (towards the telomere), successful amplification was achieved for Rep. The opposite boundary was approximated based upon the successful amplification of the DYS19 locus in all samples indicating a proximal breakpoint between 8.75 Mb (sY2180) and 10.1 Mb (DYS19). Located within this region, markers sY1078, sY1293, sY56 and SHGC-80640 generated detectable PCR fragments, narrowing down the region for this breakpoint to a 180 kb span. Fig. 3 illustrates amplification products of two representative loci, sY56, which amplified in the five AMGY deleted samples and sY605, which did not. The five phenotypically normal males lacking the AMGY gene exhibited identical PCR profiles for the battery of Y-markers examined in this high-resolution delineation of the breakpoints. This parallelism does not argue for the independent origin of these five deletions. 3.3. Assessment of Y-chromosome integrity The successful amplification of various loci distributed throughout both arms of the Y-chromosome confirmed its integrity outside the critical dropout region. Panel A of Fig. 1 depicts the positions of these markers. On the short arm, M410 and DYS393 generated positive results, while on the long arm of the Y-chromosome, amplified products were obtained for a multitude of markers DYS271, DYS391, DYS439, DYS388, M241, DYS389I/II, DYS390, DYSA7.2, YAP, M9, DYS392, and M304. These results suggest that the chromosomal alterations exhibited by the five phenotypically normal male individuals are limited and localized to the AMGY gene and immediate flanking sequences. 3.4. Evolutionary history of the deletion Based on the Y-haplogroup analysis performed, all five individuals were found to belong to the same lineage, J2b2M241. Although the Y-STR loci studied on the amelogenindeleted individuals were not identical, they exhibit striking similarities in the number of shared alleles. The five AMGY null individuals display a total of four similar haplotypes, with two individuals having identical Y-STR profiles (Table 2). Assuming a common ancestor for the five Kathmandu individuals, the microsatellite variation was employed to estimate 6.5  3.3 ky for the age of the deletion mutation. 4. Discussion Amplification utilizing amelogenin primers against target sequences outside the original Identifiler1 oligonucleotides demonstrated that the AMGY null electropherogram profiles

were not due to a PCR failure as a consequence of point mutations within the annealing region of the primers, but was caused by deletions involving the amelogenin locus on this chromosome. The absence of the sY1242, sY605, sY74, M12, MSY1, DYS257b, and sY2180 loci was observed in all five amelogenin-deleted males and suggests a deletion of about 2.3 Mb spanning from the sY1242 locus (6.443 Mb) to the sY2180 marker (8.75 Mb). The approximate size of the deletion corroborates the findings from previous AMGY null studies that define regions of 1Mb through Southern hybridization [15] and 2.5 Mb through fluorescence in situ hybridization (FISH) [18]. Successful amplification of STS markers Rep and SHGC80640 further refines the region where the breakpoints of the deletion occurred. While the distal breakpoint is restricted to approximately 5 kb between sY1242 and Rep, the proximal lies within the 180 kb between sY2180 and SHGC-80640. Similarly, previous publications have attempted to pinpoint the extent of previously detected AMGY deletions. Lattanzi et al. [18] characterized the distal breakpoint within the 11,538 bp between markers Y7567 (the reverse primer used for sY1091) and sY1242. According to this previous study, this represents a highly repetitive area where primers cannot be designed [18]. Another report analyzing a different set of amelogenin null individuals delineated the location of the proximal breakpoint to the 5 Mb between minisatellite MSY1 and DYS439 [16]. Since each study employed distinct loci, it is difficult to infer a common deletion across studies. However, the markers examined in the present report allowed chromosome walking along the proximal and distal breakpoint regions on each side (5 and 180 kb, respectively) to generate a highresolution map of the deletions. Using this approach, the five AMGY dropout males display identical regional localization for their deletions, supporting the idea of a common ancestor. In addition, the fact that the distal breakpoint reported by Lattanzi et al. [18] in two individuals collected in Italy coincides with the high-resolution mapping of this truncation in the present study may suggest that these deletions were not a result of independent mutations. Alternatively, the parallelism observed in the different studies regarding approximate site of cleavage could be explained by the existence of breakage hot spots. Although the event(s) that led to the five AMGY null Kathmandu males represent localized regional deletions, the possibility of additional chromosome aberrations need to be explored. The Y-biallelic markers analyzed in these five males were successfully detected, discarding the possibility of extensive chromosome-wide deletions. The successful amplification of YAP, M9, M241, M304, and M410, located in areas upstream and downstream of the critical dropout region suggest that the integrity of the Y-chromosome is not compromised in the probed areas in these five samples. Further support for the overall completeness of these five Y-chromosomes is provided by the positive results obtained from the 10 microsatellites examined since they are also dispersed to varying points on either arm of this chromosome. The evolutionary history of the AMGY deletions was explored through Y-haplogroup analysis. The five AMGY null individuals belong to the same haplogroup, J2b2-M241. The

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

161

males for the remaining three publications listed in Table 2 displays a greater proportion of similarity within each study. The Y-STR analysis of the five Kathmandu amelogenin null males uncovered four related haplotypes, overall sharing a minimum of seven alleles and a maximum of 10. The fact that the five samples belong to the same Y-haplogroup (J2b2-M241) and exhibit identical (as in the case of Kat28 and Kat66) or very similar Y-STR haplotypes suggests that the deletions possibly occurred in the same paternal lineage. Assuming a common ancestor for the deletions, the age of the AMGY mutation determined by microsatellite variation is 6.5  3.3 ky. Since the estimated age of M241 is 13.8  3.8 ky [23], the inferred age generated for the amelogenin deletion from the Kathmandu data is as expected younger than the total age of the J2b2 haplogroup as defined by the M241 SNP. Although interstudy comparisons are limited due to the absence of any phylogenetic information concerning haplogroup assignment and the analysis of different Y-STRs, some parallelism is observed among the AMGY null individuals reported in the four publications in Table 2. For example, nine of the eleven loci (DYS19, DYS389I, DYS389II, DYS390, DYS391, DYS392, DYS393, DYS439 and DYS438) are examined in more than one study and exhibit alleles in common among the individuals of these independent reports. Furthermore, five of those nine loci (DYS390, DYS391, DYS392, DYS439 and DYS438) possess alleles in common in all the studies that reported on that particular Y-STR. Although haplogroup data is not available from previous reports of AMGY deletions, the possibility that all individuals exhibiting the amelogenin dropouts are related by descent is thought provoking. Failure of the amelogenin test suggests that more reliable loci for sex identification should be incorporated, particularly

ancestry of J2b2 can be traced back to the SNP mutation J2M172, thought to be associated with the spread of agriculture from Anatolia [31,37]. The estimated age for haplogroup J2, 18.5  3.5 ky [37], supports its dissemination during the Neolithic as part of the agricultural revolution that occurred in the Fertile Crescent. The pattern of distribution for a subsequent mutation, J2b-M12, is consistent with a diffusion into Europe from the southern Balkans [37]. Finally the last known marker in this lineage is the J2b2 haplogroup defined by the M241 SNP site first reported by Cinniog˘lu et al. [31] in 0.96% of Turkish males. The highest frequencies of this lineage were subsequently detected in India and Pakistan (5.22% and 2.27%, respectively) [23] as well as in the Nepali populations of Kathmandu (6.49%) and Newar (1.5%) (Gayden, unpublished results). These distributions suggest that the genesis of M241 may reside within or near the Indian subcontinent. This idea is corroborated by the regional concentration of afflicted individuals reported in previous publications. The percent of amelogenin dropouts in the Kathmandu population (6.49%) is lower than that observed by Santos et al. [13] in Sri Lankan males (8%) but higher than that of males observed in India (3.57%) and Malaysia (0.88%) by Chang et al. [16], India (1.85%) by Thangaraj et al. [15], Israel (1.04%) by Michael and Brauner [17], Austria (0.018%) by Steinlechner et al. [14] and Italy (0.008%) by Lattanzi et al. [18]. Further localization of the possible geographic origin of the J2b2 AMGY deleted haplogroup will require the identification of where the average Y-STR diversity is highest with lower Y-STR diversities occurring towards the perimeter of the distribution [23]. With the exception of the work by Lattanzi et al. [18] in which both AMGY null individuals possess very different YSTR haplotypes, comparisons of the Y-STR profiles of affected Table 2 Y-STR haplotypes Individual

Origin

DYS 19

DYS 388

DYS 389I

DYS 389II

DYS 390

DYS 391

DYS 392

DYS 393

DYS 439

DYSA 7.2

Kat16 Kat28 Kat50 Kat62 Kat66

Nepal Nepal Nepal Nepal Nepal

15 15 13 15 15

No amp 15 15 15 15

12 12 12 12 12

29 29 29 29 29

23 23 23 24 23

10 10 9 10 10

11 11 11 11 11

12 12 12 12 12

12 13 12 12 13

7 7 7 7 7

CA1 W121 W116 W140 AI25

Malay Indian Indian Indian Indian

R2 M60 B16 K96 A151

Indian Indian Indian Indian Indian

14 14 16 17 14

13b 13b 14b 13b 13b

30 b 30 b 30 b 30 b 30 b

22 22 24 25 21

10 10 11 10 11

710/01 121/04

c

16 13

14 13

30 29

22 24

11 11

a b c

c

25 25 24 25 23

13 11 13 12 13

DYS 438

10 9 9 9 9

14 14 12 12 14 11 13

14 13

Comparisons made between individuals within each study. Constant three repeats (3 TCTG) included according to new nomenclature (ystr.org). Individuals of unknown biogeographical origin.

11 12

10 12

Max/min shared allelesa

Reference

8/7 10/7 7/7 8/7 10/7

Present study

1/1 2/1 2/1 2/1 2/1

Chang et al. [16]

6/1 6/1 2/1 3/2 4/2

Thangaraj et al. [15]

1/1 1/1

Lattanzi et al. [18]

162

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163

for populations where the frequency of the deletion is higher and gender misidentification can occur as a result. Based on the populations affected in this and previous studies it seems that the AMGY deletions are more prevalent in a distinct region of the world, specifically the Indian subcontinent. Although it is possible to infer a common ancestor as the source of the deletion in the five Kathmandu males based on the combined data from the present study, little information exists to assess whether the deletions observed in other publications are related by descent or arose from independent mutations. In light of the considerable progress made in the past decade in defining and refining the basic cladistic framework of the global Ychromosome phylogeny, a comprehensive analysis of additional individuals with AMGY deleted chromosomes utilizing Y-STRs and Y-chromosome biallelic polymorphisms to elucidate the degree of phylogenetic haplogroup and haplotype affinities warrants investigation. Note added in proof A recent report of amelogenin deletions in haplogroup J2M172 chromomosomes in Indian populations (Kashyap et al., 2006) is consistent with our more phylogenetically resolved haplogroup J2b2-M241 results following nomenclature of Sengupta et al. [23]. Acknowledgements The authors would like to thank Karen Tellez-Jacques for technical assistance provided during this study. References [1] E.C. Lau, T.K. Mohandas, L.J. Shapiro, H.C. Slavkin, M.L. Snead, Human and mouse amelogenin gene loci are on the sex chromosomes, Genomics 4 (1989) 162–168. [2] E.C. Salido, P.H. Yen, K. Koprivnikar, L. Yu, L.J. Shapiro, The human enamel protein gene amelogenin is expressed from both the X and the Y chromosomes, Am. J. Hum. Genet. 50 (1992) 303–316. [3] A. Akane, H. Shiono, K. Matsubara, Y. Nakahori, S. Seki, S. Nagafuchi, M. Yamada, Y. Nakagome, Sex identification of forensic specimens by polymerase chain reaction (PCR): two alternative methods, Forensic Sci. Int. 49 (1991) 81–88. [4] Y. Nakahori, O. Takenaka, Y. Nakagome, A human X–Y homologous region encodes amelogenin, Genomics 9 (1991) 264–269. [5] Y. Nakahori, K. Hamano, M. Iwaya, Y. Nakagome, Sex identification by the polymerase chain reaction using X–Y homologous primers, Am. J. Med. Genet. 39 (1991) 472–473. [6] K.M. Sullivan, A. Mannucci, C.P. Kimpton, P. Gill, A rapid and quantitative DNA sex test: fluorescence-based PCR analysis of X–Y homologous gene amelogenin, Biotechniques 15 (1993) 636–641. [7] S. Repping, C.M. Korver, R.D. Oates, S. Silber, F. van der Veen, D.C. Page, S. Rozen, Are sequence family variants useful for identifying deletions in the human Y-chromosome? Am. J. Hum. Genet. 75 (3) (2004) 514–517. [8] P.H. Vogt, AZF deletions and Y chromosomal haplogroups: history and update based on sequence, Hum. Reprod. Update 11 (4) (2005) 319–336. [9] S. Fernandes, S. Paracchini, L.H. Meyer, G. Floridia, C. Tyler-Smith, P.H. Vogt, A large AZFc deletion removes DAZ3/DAZ4 and nearby genes from men in Y haplogroup N, Am. J. Hum. Genet. 74 (2004) 180–187.

[10] P.E. Roffey, C.I. Eckhoff, J.L. Kuhl, A rare mutation in the amelogenin gene and its potential investigative ramifications, J. Forensic Sci. 45 (2000) 1016–1019. [11] J.G. Shewale, S.L. Richey, S.K. Sinha, Anomalous amplification of the amelogenin locus typed by AmpFlSTR1 Profiler PlusTM Amplification Kit, Forensic Sci. Commun. 2 (4) (2000). [12] B. Shadrach, M. Commane, C. Hren, I. Warshawsky, A rare mutation in the primer binding region of the amelogenin gene can interfere with gender identification, J. Mol. Diagn. 6 (4) (2004) 401–405. [13] F.R. Santos, A. Pandya, C. Tyler-Smith, Reliability of DNA-based sex tests, Nat. Genet. 18 (1998) 103. [14] M. Steinlechner, B. Berger, H. Niederstatter, W. Parson, Rare failures in the amelogenin sex test, Int. J. Legal Med. 116 (2002) 117–120. [15] K. Thangaraj, A.G. Reddy, L. Singh, Is the amelogenin gene reliable for gender identification in forensic casework and prenatal diagnosis? Int. J. Legal Med. 116 (2002) 121–123. [16] Y.M. Chang, L.A. Burgoyne, K. Both, Higher failures of amelogenin sex test in an Indian population group, J. Forensic Sci. 48 (2003) 1309–1313. [17] A. Michael, P. Brauner, Erroneous gender identification by the amelogenin sex test, J. Forensic Sci. 49 (2) (2004) 258–259. [18] W. Lattanzi, M.C. Di Giacomo, G.M. Lenato, G. Chimienti, G. Voglino, N. Resta, G. Pepe, G. Guanti, A large interstitial deletion encompassing the amelogenin gene on the short arm of the Y chromosome, Hum. Genet. 116 (5) (2005) 395–401. [19] B. Brinkmann, Is the amelogenin sex test valid? Int. J. Legal Med. 116 (2002) 63. [20] K. Muroya, T. Ishii, Y. Nakahori, Y. Asakura, K. Tachibana, M. Masuno, K. Imaizumi, Y. Tanaka, Y. Kawada, S. Yukizane, T. Ogata, Gonadoblastoma, mixed germ cell tumor, and Y chromosomal genotype: molecular analysis in four patients, Genes Chromosomes Cancer 25 (1999) 40–45. [21] L. Tiepolo, O. Zuffardi, Localization of factors controlling spermatogenesis in the nonfluorescent portion of the human Y chromosome long arm, Hum. Genet. 34 (1976) 119–124. [22] G. Antunez de Mayolo, A. Antunez de Mayolo, P. Antunez de Mayolo, S.S. Papiha, M. Hammer, J.J. Yunis, E.J. Yunis, C. Damodaran, M. Martinez de Pancorbo, J.L. Caeiro, V.P. Puzyrev, R.J. Herrera, Phylogenetics of worldwide human populations as determined by polymorphic Alu insertions, Electrophoresis 23 (2002) 3346–3356. [23] S. Sengupta, L.A. Zhivotovsky, R. King, S.Q. Mehdi, C.A. Edmonds, C.T. Chow, A.A. Lin, M. Mitra, S.K. Sil, A. Ramesh, M.V. Usha Rani, C.M. Thakur, L.L. Cavalli-Sforza, P.P. Majumder, P.A. Underhill, Polarity and temporality of high-resolution Y-chromosome distributions in India identify both indigenous and exogenous expansions and reveal minor genetic influence of Central Asian pastoralists, Am. J. Hum. Genet. 78 (2) (2006) 202–221. [24] C.A. Tilford, T. Kuroda-Kawaguchi, H. Skaletsky, S. Rozen, L.G. Brown, M. Rosenberg, J.D. McPherson, K. Wylie, M. Sekhon, T.A. Kucaba, R.H. Waterston, D.C. Page, A physical map of the human Y chromosome, Nature 409 (6822) (2001) 943–945. [25] D. Vollrath, S. Foote, A. Hilton, L.G. Brown, P. Beer-Romero, J.S. Bogan, D.C. Page, The human Y chromosome: a 43-interval map based on naturally occurring deletions, Science 258 (5079) (1992) 52– 59. [26] P.A. Underhill, L. Jin, A.A. Lin, S.Q. Mehdi, T. Jenkins, D. Vollrath, R.W. Davis, L.L. Cavalli-Sforza, P.J. Oefner, Detection of numerous Y chromosome biallelic polymorphisms by denaturing high-performance liquid chromatography, Genome Res. 7 (10) (1997) 947–949. [27] J.E. Sulston, R. Waterston, Toward a complete human genome sequence, Genome Res. 8 (11) (1998) 1097–1108. [28] T. Hudson, Whitehead Institute/MIT Center for Genome Research; Physically Mapped STSs, 1995, unpublished. [29] T. Kuroda-Kawaguchi, H. Skaletsky, P.J. Minx, L.G. Brown, S. Rozen, R.K. Wilson, R.H. Waterston, D.C. Page, The DNA sequence of the human Y chromosome, unpublished. [30] M.T. Seielstad, J.M. Hebert, A.A. Lin, P.A. Underhill, M. Ibrahim, D. Vollrath, L.L. Cavalli-Sforza, Construction of human Y-chromosomal haplotypes using a new polymorphic A to G transition, Hum. Mol. Genet. 3 (12) (1994) 2159–2161.

A.M. Cadenas et al. / Forensic Science International 166 (2007) 155–163 [31] C. Cinniog˘lu, R. King, T. Kivisild, E. Kalfoglu, S. Atasoy, G.L. Cavalleri, A.S. Lillie, C.C. Roseman, A.A. Lin, K. Prince, P.J. Oefner, P. Shen, O. Semino, L.L. Cavalli-Sforza, P.A. Underhill, Excavating Y-chromosome haplotype strata in Anatolia, Hum. Genet. 114 (2004) 127–148. [32] M.F. Hammer, S. Horai, Y chromosomal DNA variation and the peopling of Japan, Am. J. Hum. Genet. 56 (1995) 951–962. [33] Y Chromosome Consortium, A nomenclature system for the tree of human Y-chromosomal binary haplogroups, Genome Res. 12 (2002) 339–348. [34] L.A. Zhivotovsky, N.A. Rosenberg, M.W. Feldman, Features of evolution and expansion of modern humans inferred from genomewide microsatellite markers, Am. J. Hum. Genet. 72 (2003) 1171–1186. [35] L.A. Zhivotovsky, P.A. Underhill, C. Cinniog˘lu, M. Kayser, B. Morar, T. Kivisild, R. Scozzari, F. Cruciani, G. Destro-Bisol, G. Spedini, G.K.

163

Chambers, R.J. Herrera, K.K. Yong, D. Gresham, I. Tournev, M.W. Feldman, L. Kalaydjieva, The effective mutation rate at Y chromosome short tandem repeats, with application to human population-divergence time, Am. J. Hum. Genet. 74 (2004) 50–61. [36] L.A. Zhivotovsky, P.A. Underhill, On the evolutionary mutation rate at Ychromosome STRs: comments on paper by Di Giacomo et al. (2004), Hum. Genet. 116 (2005) 529–532. [37] O. Semino, C. Magri, G. Benuzzi, A.A. Lin, N. Al-Zahery, V. Battaglia, L. Maccioni, C. Triantaphyllidis, P. Shen, P.J. Oefner, L.A. Zhivotovsky, R. King, A. Torroni, L.L. Cavalli-Sforza, P.A. Underhill, A.S. SantachiaraBenerecetti, Origin, diffusion, and differentiation of Y-chromosome haplogroups E and J: inferences on the neolithization of Europe and later migratory events in the Mediterranean area, Am. J. Hum. Genet. 74 (2004) 1023–1034.

Forensic Science International 166 (2007) 164–175 www.elsevier.com/locate/forsciint

Generating population data for the EMPOP database—An overview of the mtDNA sequencing and data evaluation processes considering 273 Austrian control region sequences as example Anita Brandsta¨tter, Harald Niedersta¨tter, Marion Pavlic, Petra Grubwieser, Walther Parson * Institute of Legal Medicine, Innsbruck Medical University, Mu¨llerstr. 44, 6020 Innsbruck, Austria Received 27 February 2006; received in revised form 9 May 2006; accepted 9 May 2006 Available online 7 July 2006

Abstract The European DNA profiling group (EDNAP) mtDNA population database (EMPOP) is an international collaborative project between DNA laboratories performing mtDNA analysis and the DNA laboratory of the Institute of Legal Medicine (GMI) in Innsbruck, Austria. The goal is to set up a directly accessible mtDNA population database, which can be used in routine forensic casework for frequency investigations. Here we describe a safe laboratory scheme involving electronical data handling and computer-aided data transfer, which help to minimize errors originating from potential sample mix-up, data misinterpretation and incorrect transcription. The procedure is demonstrated by example of an mtDNA control region population study on 273 unrelated individuals from Austria. Our population sample was compared with five other European populations via an analysis of molecular variance (AMOVA). The inclusion of regions outside HVS-I and HVS-II increased the amount of information on the haplogroup diagnostic sites in the control region. Most of the haplotypes in Austrians fell into haplogroups H, J, K, T, and U. The random match probability in Austrians was 1:125; the average number of nucleotide differences between individuals in the Austrian database was 9.32. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Mitochondrial DNA polymorphisms; Haplogroup; Phylogenetic; Mean pairwise distances; AMOVA; FST values; Random match probabilities

1. Introduction Forensic molecular biology takes advantage of the high copy number of mitochondrial DNA (mtDNA) molecules in a cell, and meanwhile, mtDNA typing has become routine in its application to analyze samples where the amount of genomic DNA is very small or degraded. Because of the lack of recombination, mtDNA profiles are not unique but regarded as haplotypes. When an mtDNA haplotype derived from an evidence sample cannot be discriminated from the one of a reference sample (e.g. from the suspect), its relative rarity can be estimated by comparing it to a collection of mtDNA sequences – usually assembled in databases – in order to support assessment of the weight to the evidence [1]. Some of the published mtDNA data have faced severe criticism in the past, as an unacceptable high rate of errors was detected therein (e.g. [2–4]). Although good laboratory practice guidelines have been published [5,6], problems still occur in the

* Corresponding author. Tel.: +43 512 5073303; fax: +43 512 5072764. E-mail address: [email protected] (W. Parson). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.006

course of mtDNA data generation. Sources of these errors are meanwhile well known and have been described in some detail [2,7,8], however, the amplification and sequencing process as well as a posteriori data control and transfer of haplotypes still bear shortcomings which result in erroneous data. We here present in detail a reliable laboratory concept for mtDNA typing based on the generation of redundant sequence information for unequivocal base-assignment, independent double data evaluation, and IT-based (manual-free) comparison and transfer of results. This newly developed strategy is showcased on 273 Austrian mtDNA control region sequences. The thus created population data will be incorporated in the EMPOP database (http://www.empop.org/) [7,9]. 2. Materials and methods 2.1. DNA samples and extraction DNA was extracted from blood samples obtained from 273 unrelated West-Eurasians from Austria using Chelex 100 as outlined in [10].

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

2.2. Amplification and sequencing of the entire mtDNA control region

Table 1 Sequences of primers Primer

In order to facilitate efficient sequencing of the mitochondrial DNA control region, a 96-well method of processing population samples was developed independently by two laboratories: AFDIL, Rockville, USA [11–13] and GMI Innsbruck, Austria [14]. In the latter aliquots of DNA extracts were dispensed in 96-well plates; four negative control samples were assayed per plate. For amplification, to each well of a MicroAmp Optical 96-well reaction plate (AB), 18 ml of PCR master mix containing 1.0 unit of AmpliTAQ Gold polymerase (AB, Foster City, CA), 1.0 unit of PCR reaction buffer (AB), 200 mM each dNTP (AB), and 0.5 mM each primer (L15900 and H00599) were dispensed. Then 2 ml DNA extract were put into the PCR master mix with a multichannel pipette. The amplification reaction was conducted on a thermal cycler (e.g. 9600/9700 GeneAmp Thermal Cycler; AB). The reaction cocktails were heated to 95 8C (11 min) and then put through 35 reaction cycles: 95 8C for 15 s, 56 8C for 30 s and 72 8C for 90 s, followed by a final extension phase at 72 8C for 10 min. PCR primers and unincorporated dNTPs were removed by adding 8 ml of ExoSAP-IT (USB, Cleveland, OH) with a repeater pipette (Eppendorf Multipette; Hamburg, Germany) and heating the samples to 37 8C (15 min) for enzyme activation and then to 80 8C (15 min) for enzyme deactivation. Subsequently, 2 ml of purified PCR product were combined with the sequencing master mix (containing 2 ml BigDye Terminator v1.1 Cycle Sequencing RR mix (AB), 2 ml BigDye Terminator v1.1 Sequencing Buffer (AB), 1.6 pmol primer and distilled water up to 8 ml) with an 8-channel epMotion workstation (Eppendorf, Germany). Cycle sequencing was performed (after a first denaturation step of 95 8C, 1 min) for 25 cycles of 10 s at 95 8C, 5 s at 50 8C, and 4 min at 60 8C. Each template was sequenced in the forward direction with primers L15971, L15989, L16268, L00015, L00029, L00314, and L00361 and in the reverse direction with primers H00016, H00159, and H00484 (Table 1(a) and (b)). Sequencing reaction products were purified from residual dye terminators using Sephadex G-50 Fine (Amersham, Buckinghamshire, UK) and Multiscreen filter plates (Millipore) according to the manufacturer’s protocol. To this end, the cycle sequencing products were diluted by adding 10 ml of distilled water and the dilutions were centrifuged through the filter plate into an optical 96well plate for electrophoretic separation. The entire procedure of diluting cycle sequencing products and transferring the dilutions onto the Sephadex columns in the filter plate was again performed by the epMotion workstation. When spinning cycle sequencing products through the filter plate, unequal amounts of product may be recovered throughout the plate. In order to avoid this, the blocks’ orientations in the centrifuge carriage were reversed after 2.5 min and the blocks were spun a second time for 2.5 min to obtain consistent amounts of purified products. Electrophoretic separation was carried out on an ABI3100 capillary sequencer using POP6 and a 36 cm capillary array.

165

Nucleotide sequence

(a) Sequences of primers used for amplification and sequencing of the entire mitochondrial DNA control region H00016 50 TGA TAG ACC TGT GAT CCA TCG TGA 30 H00159 50 AAA TAA TAG GAT GAG GCA GGA ATC 30 H00484 50 TGA GAT TAG TAG TAT GGG AG 30 H00599 50 TTG AGG AGG TAA GCT ACA TA 30 L00015 50 CAC CCT ATT AAC CAC TCA CG 30 L00029 50 CTC ACG GGA GCT CTC CAT GC 30 L00314 50 CCG CTT CTG GCC ACA GCA CT 30 L00361 50 ACA AAG AAC CCT AAC ACC AGC 30 L15900 50 TCA AAG CTT ACA CCA GTC TTG TAA ACC 30 L15971 50 TTA ACT CCA CCA TTA GCA CC 30 L15989 50 CCC AAA GCT AAG ATT CTA AT 30 L16268 50 CAC TAG GAT ACC AAC AAA CC 30 (b) Sequencing primers that were published in [14] and that were replaced in the present studya L16169 50 CCC CCC CCC CAT G 30 (replaced by L00029) L00318 50 CCC CCC CCC CCC GCT 30 (replaced by L00314) L00403 50 TCT TTT GGC GGT ATG CAC TTT 30 (replaced by L00361) H16410 50 GAG GAT GGT GGT CAA GGG AC 30 (replaced by H00016) a

The primers, which they were substituted for, are written in parentheses.

2.3. Data analysis and quality assurance After the raw data were analyzed with Sequencing Analysis (Version 3.3, Applied Biosystems, AB, Foster City, CA, USA), the sequences were aligned and the basecalls were scrutinized twice by two independent scientists. Consensus sequences were aligned and compared to the revised Cambridge Reference Sequence (rCRS; [15,16]) using Sequencher (Version 4.1.4Fb4, GeneCodes, Ann Arbor, MI, USA), following nomenclature guidelines for mtDNA typing [5,6]. In an independent step the sequences were evaluated by another scientist using the sequence analysis and alignment software SeqScape (Version 2.0, AB). The results of both analyses were exported as mutation reports, which were directly imported into a data storage and data evaluation software, specifically designed for mtDNA analysis and implemented into the previously described in-house Laboratory Information Management System (LIMS) ([17]). This software allows for electronical comparison of the exported haplotypes, which are finally reviewed by a third scientist. After evaluation the mtDNA haplotypes were stored and assembled for subsequent phylogenetic analyses and final database export. 2.4. Haplogroup assignment The mtDNA haplotypes from this study were affiliated to (sub)-haplogroups based on the patterns of shared haplogroup-

166

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

specific or haplogroup-associated polymorphisms in the control region, as reported in [18–27]. The haplogroup assignments were confirmed with the results of a previous study involving the analysis of 16 phylogenetic informative single nucleotide polymorphisms (SNPs) from the coding region [28]. 2.5. Random match probability The random match probability was calculated as the sum of the squares of the haplotype frequencies [29]. C-stretch length variants in HVS-I (around position 16189) and in HVS-II (around position 310) were ignored in distinguishing haplotypes for calculation of random match probability.

the mtDNA haplotypes were generated by sequence analysis of the full control region amplicons. Therefore, the risk of mixing up sample was significantly reduced. Nevertheless, we tested the haplotypes for artificial recombination, which could have potentially been introduced at a later stage of the analysis. The second method applied to the data was a mathematical algorithm developed to aid the determination of the most probable haplogroup(s) of a set of mtDNA control region samples. The algorithm is implemented into a software package (manuscript in preparation) and is based on propositional logic via checking of the presence or absence of haplogroup diagnostic sites. 3. Results and discussion

2.6. Population genetic and molecular evolution analysis The Austrian data were compared with 93 control region (CR) sequences from the Czech Republic [30], 244 CR sequences from Poland [19], 156 CR sequences from Bosnia-Herzegovina [22], 104 CR sequences from Slovenia [22] and 200 CR sequences from Germany [31]. For all comparisons, C-stretch length variants in HVS-I and HVS-II were ignored. Molecular diversity indices, analysis of molecular variance (AMOVA) and pairwise differences were calculated with the ARLEQUIN software (Version 2.0; [32]). Pairwise FST values were used to describe the short-term genetic distance between populations. Permutation tests (1000 replicates) were used to evaluate the significance of calculated genetic distances between populations. For this alignment, sequences were trimmed to fit the greatest common range 16024–16368 and 71–340. 2.7. Phylogenetic analyses Two different kinds of phylogenetic analyses were performed on the data. The first method was aimed at identifying possible artificial HVS-I/HVS-II recombinants (i.e. samples that contain a combination of HVS-I sequences from one person and HVS-II sequences from another person). The occurrence of artificial recombination in mtDNA databases has been described several times [33,3,7], and is difficult to address as this phenomenon cannot be detected by evaluation of the raw sequence data. To address this problem, we have developed an IT-based solution to systematically search for artificial recombinants in population data that consist of HVSI + HVS-II-sequences. In brief, the computer program identifies the mitochondrial haplogroup of each hypervariable region of a sample separately. This is achieved by finding a haplotype with the minimal genetic distance to the sequence in question from a flexible background dataset containing control region sequences, whose haplogroup-affiliation was confirmed by coding region SNPs. The distance is calculated applying a combination of the simple Hamming-distance for sites with gaps (insertions/deletions) with the more complex GTR (general time reversible; [34]) distance for sites without gaps. Then, the implementation checks whether the hg-assignments of the two hypervariable regions are conclusive—if not, it is indicated to re-examine the mismatching profile. In this study,

We here present an optimized laboratory strategy for the sequence analysis of the complete mitochondrial DNA control region, which involves semi-automated 96 well-based pipetting and LIMS-and IT-aided sample processing. A sample of 273 unrelated persons from Austria serves as example (Table 2). This strategy developed for high-throughput typing of mtDNA control region sequences has proven to be a reliable and easily reproducible procedure for generating high-quality population data. The concept of keeping all laboratory stages in the 96-well format facilitates the handling of the samples and minimizes the possibility of artificial recombination (sample mix-up of hypervariable segments from different individuals [35,3,7]) considerably. Another important step in the avoidance of sample mix-up is that the entire control region is amplified in one piece, which reduces the complexity of post-PCR purification treatment and the further transfer of PCR products into the different cycle sequencing cocktails. The choice of the number of PCR cycles (35) is based on the intention to amplify the samples until the plateau phase of PCR is reached, largely independent of their initial DNA concentrations. This normalizes the amount of DNA that is finally added to the cycle sequencing reaction. Also in the post-PCR area, the application of automated pipetting workstations for transferring PCR-products into the different cycle sequencing cocktails, diluting the cycle sequencing products and transferring the diluted cycle sequencing products onto the Sephadex columns in the Multiscreen filter plate, enables a fast and reliable processing of a large number of samples by significantly reducing the number of manual pipetting steps. The sequencing reactions produce widely overlapping DNA sequences, generating overlapping redundant sequence information across the entire mitochondrial control region. This has proven very useful for the determination of point heteroplasmic positions as these were usually confirmed multiple times by independent sequence strands. Even in the case of length heteroplasmy occurring in any of the three C-stretches within the control region, at least full double strand coverage of all nucleotide positions was given with the 10 sequencing primers (Table 1(a)). The major benefit of this strategy is to reduce the need for repeated individual data handling due to length heteroplasmy or artifacts, which challenge sequence interpretation and require additional analyses. Some of the

Table 2 Control region sequences in Austria (haplogroup-assignments were confirmed with coding region SNPs [28]) HVS-I (16024–16569)

f1G6 f1A5 f1A6 f1B3 f1B6 f1C1 f1D1 f1E1 f1E2 f1E3 f1F1 f1G2 f1G3 f1H2 f1H3 f1H5 f2A4 f2B1 f2B5 f2D1 f2D3 f2E4 f2E5 f2F3 f2F5 f2G1 f2G2 f2H4 f3A5 f3A6 f3B1 f3B4 f3B5 f3B6 f3C2 f3C3 f3C4 f3D5 f3E1 f3E4 f3F3 m1A2 m1A4 m1A5 m1B3 m1B4 m1D2 m1D5 m1E2 m1E3 m1E5 m1F1 m1F2 m1F3 m1F5 m1H1 m2B2 m2B3 m2B6 m2C2

16093C 16519C 16037G

16223T

16234T

16188A

16519C

16248T 16519C 16183C 16293G 16291T 16093C 16234T 16311C 16325C 16519C 16311C 16266T 16189C 16189C 16189C 16519C 16072T 16519C 16519C 16293G 16519C 16311C

16519C

16172Y 16519C 16189C 16519C 16519C 16248T 16519C 16209C 16183C

16327T

16519C 16093Y 16093C 16299G 16519C 16239T 16519C 16234T 16129A 16519C 16519C 16519C 16519C 16519C 16170G 16327T 16291T 16042A 16261T 16519C 16124C

16189C 16311C 16519C 16271C 16293G

HVS-II (1–576) 16288C

16519C

16519C 16519C

16519C 16519C 16311C 16519C 16519C 16519C

16399G

16245T

16311C

16294T

16311C

16519C

16519C

16519C

16519C

16519C 16189C

16519C

16519C 16519C 16519C 16355Y 16527T 16293G 16311C

16390A

16288C 16291T 16519C

16519C 16519C 16519C

16519C

16519C 16311C

16519C

16298C

16327T

16518T

16519C

16527T

73G 235G 263G 263G 73G 189G 263G 143A 263G 263G 263G 263G 263G 153G 263G 263G 263G 152C 263G 146C 263G 263G 263G 143A 152C 152C 263G 263G 146C 152C 152C 259G 146C 263G 263G 146C 263G 263G 309.1C 263G 263G 263G 150T 263G 263G 146C 263G 195C 263G 263G 263G 146C 152C 263G 263G 193G 263G 200G 263G 263G

249del 263G 309.1C 315.1C 187T 263G 309.1C 195C 315.1C 309.1C 315.1C 309.1C 309.1C 263G 309.1C 315.1C 315.1C 263G 315.1C 263G 315.1C 309.1C 309.1C 195C 263G 263G 309.1C 309.1C 263G 263G 263G 263G 263G 309.1C 315.1C 263G 315.1C 309.1C 315.1C 309.1C 309.1C 315.1C 263G 309.1C 309.1C 195C 315.1C 263G 309.1C 309.1C 309.1C 263G 263G 309.1C 309.1C 263G 315.1C 263G 309.1C 309.1C

263G 309.1C 309.2C 340T 263G 315.1C 309.2C 263G 315.1C 315.2C 315.1C 315.1C 309.1C 315.1C 523del

309.1C 309.2C 315.1C 523del 315.1C

315.1C 315.1C 524.1A 524del 523del

315.1C 309.1C

315.1C

523del

524del

524.2C 524del

315.1C 524del

309.1C

315.1C

309.1C 523del 315.1C 315.1C 263G 309.1C 315.1C 309.2C 315.1C 309.1C 309.1C 315.1C 315.1C 309.1C 309.2C

315.1C 524del

309.2C 315.1C

315.1C

309.1C 523del 309.2C

309.2C 524del 315.1C

315.1C

315.1C 315.1C

573.1-6CCCCCC

315.1C 309.2C 315.1C 263G 315.2C 309.1C 315.1C 315.1C 315.1C 309.1C 315.1C 315.1C 315.1C 315.1C 477C 309.1C 315.1C 315.1C

489C

385G

309.1C 315.1C

315.1C 334C

524.1A

315.1C 438T 315.1C 309.2C 524.1A

523del

524del

315.1C 524.2C

385G

524.2C

315.1C 315.1C 315.1C

309.2C

315.1C

438T

315.1C 573.1-3CCC

167

Sample

C H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H*

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

Hg

168

Table 2 (Continued ) Sample

HVS-I (16024–16569)

HVS-II (1–576)

H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H* H10 H1a H1a H1a H1a H1a H1b H1b H1b H1b H1b H1c H1c H1c H1c H1c1 H1c1 H2a1 H5 H5 H5 H5 H5 H5 H5 H5 H5 H6 H6 H6 H6 H6 H6 H6 H6 HV* HV0

m2C4 m2C6 m2D3 m2E4 m2E6 m2F1 m2F3 m2F4 m2G1 m2H6 m3A4 m3A5 m3C2 m3D3 m3E4 m3E5 m3F1 m3F6 m3G1 m3G3 m3G5 m3G6 m3H1 m3H4 m3H5 m3B6 f2C4 f3G3 f3G4 m1C2 m2D2 f1D4 f1E4 f2C6 f2E1 m2D4 f1C2 f1C5 f1C6 m3C5 f3C6 m3E6 f2D5 f1C3 f1F5 f1G4 f2E3 f3D1 m1H5 m2A4 m2B1 m3B1 f3D6 f3F1 f3H1 m1A1 m1F4 m1G5 m2D1 m3E1 m3G4 f1G1

16327T 16519C 16129A 16519C 16519C 16239T 16249C 16519C 16170G 16519C 16183C 16183C 16311C 16519C

263G 263G 263G 263G 152C 263G 263G 146C 263G 263G 263G 199C 263G 263G 73G 152C 263G 199C 72G 152C 246C 152C 263G 263G 262T 263G 73G 73G 73G 73G 73G 263G 263G 263G 263G 263G 263G 263G 263G 263G 263G 263G 152C 146C 263G 263G 152C 263G 263G 263G 199C 263G 42.1T 239C 239C 239C 152C 183G 239C 239C 263G 72C

16519C

16311C 16519C 16527T 16355T

16390A

16189C 16189C

16519C 16519C

16207G

16519C

16183C 16519C 16519C 16093C 16216R 16256T 16249C

16189C

16519C 16311C 16311C 16519C

16114T 16162G 16051G 16162G 16162G 16051G 16189C 16189C 16189C 16189C 16183C 16093C

16519C 16209C 16162G 16519C 16519C 16162G 16356C 16356C 16356C 16356C 16189C 16263C

16278Y 16519C 16263C 16093C 16354T 16304C 16260T 16304C 16304C 16304C 16189C 16260T 16304C 16304C 16362C 16311C 16362C 16311C 16362C 16362C 16298C 16362C 16162G 16153A

16519C 16519C 16263C

16519C

16519C

16519C 16519C

16519C 16291T

16304C

16519C

16291T 16362C 16362C 16362C 16362C 16356C 16311C

16304C 16519C 16519C 16519C 16519C 16362C 16390A

16519C

16519C

16304C

16319A 16304C 16304C 16311C 16482G 16362C 16362C 16482G 16482G 16362C 16482G 16298C 16298C

16519C

16482G 16482G

16482G 16526A

16519C 16519C

309.1C 315.1C 309.1C 309.1C 263G 315.1C 315.1C 195C 315.1C 309.1C 309.1C 263G 309.1C 315.1C 182T 263G 315.1C 263G 263G 263G 263G 263G 315.1C 309.1C 263G 309.1C 263G 263G 263G 263G 263G 309.1C 309.1C 315.1C 309.1C 315.1C 315.1C 309.1C 315.1C 315.1C 315.1C 315.1C 194T 195C 309.1C 309.1C 263G 309.1C 309.1C 309.1C 207A 315.1C 239C 263G 263G 263G 239C 239C 263G 263G 309.1C 93G

315.1C

438T

309.2C 315.1C 315.1C

315.1C 523del

524del

263G 478G 315.1C 309.2C 309.1C 315.1C

315.1C 523del

524del

315.1C 309.2C

315.1C

263G 315.1C 523del 309.1C 315.1C 309.1C 309.1C 309.1C

309.1C

315.1C

524del 309.2C

315.1C

315.1C 309.2C 309.2C

315.1C 315.1C

315.1C 309.1C 315.1C 315.1C 315.1C 315.1C 315.1C 315.1C 315.1C 315.1C 523del 309.2C 523del 477C 315.1C 477C 477C 477C 477C 263G 263G 315.1C 309.2C 315.1C 309.2C 309.2C 315.1C 263G 456T 263G 309.1C 309.1C 309.1C 263G 263G 309.1C 309.1C 315.1C 195C

523del

524del

315.1C

515G

523del 523del 524del 315.1C 524del

524del 524del 523del

524del

309.1C 309.1C 456T 315.1C 456T 315.1C 315.1C 456T 315.1C

315.1C 315.1C

456T

309.1C 309.2C 309.2C 315.1C 309.1C 315.1C 315.1C 315.1C

315.1C 315.1C 315.1C

263G

315.1C

477C

456T 456T 456T

524.1A

456T

309.2C

315.1C

524.2C

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

Hg

16298C 16183C 16298C 16216G 16145A 16129A 16086C 16069T 16063C 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16069T 16048A 16224C 16224C 16224C 16093C 16224C 16224C 16224C 16222T 16224C 16093C 16093C 16224C 16224C 16224C 16224C 16224C 16224C 16224C 16214T 16224C 16224C 16224C 16093C 16147A 16086C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C

16311Y 16189C 16261T 16298C 16148T 16129A 16126C 16069T 16126C 16093C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16069T 16311C 16298C 16311C 16189A 16311C 16311C 16287T 16224C 16257T 16224C 16224C 16311C 16311C 16311C 16311C 16311C 16311C 16311C 16224C 16233G 16239T 16301Y 16224C 16154C 16147A 16294T 16172C 16266T 16153A 16241G 16294T 16294T 16163G 16163G 16163G

16298C 16298C 16519C 16192T 16223T 16193T 16086Y 16193T 16126C 16324C 16261T 16201T 16519C 16261T 16201T

16519C 16223T 16268T 16519C 16126C 16319A 16311C

16294T 16391A

16366T

16390A

16265G

16319A

16232S

16265G

16519C 16519C

16168Y 16519C

16519C

16319A

16390A 16519C

16145A 16145A 16145A 16126C 16519C 16311C 16519C 16224C 16519C 16519C 16311C 16311C 16311C 16311C 16311C 16362Y 16519C 16519C 16519C 16519C 16519C 16320T 16311C 16311C 16311C 16311C 16311C 16172C 16223T 16304C 16294T 16294T 16258G 16294T 16304C 16311C 16186T 16186T 16186T

16231C 16231C 16231C 16193T

16261T 16261T 16261T 16356C

16519C 16274A

16519C 16362C 16519C 16362C 16519C 16519C

16519C 16320T 16519C 16519C 16519C 16319A 16223T 16248T 16519C 16304C 16304C 16294T 16304C 16519C 16519C 16189C 16189C 16189C

16311C

16362C

16519C

16519C 16519C 16524G

16519C

16463G 16320T 16320T

16519C 16355T 16355T

16519C 16519C 16519C 16519C

16294T 16294T 16294T

16519C 16519C 16519C

16519C 16519C

64T 72C 72C 72C 72C 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G

72C 195C 152C 263G 263G 153G 152C 152C 263G 152C 185A 185A 185A 185A 185A 185A 185A 146C 185A 185A 185A 146C 185A 228A 185A 185A 150T 150T 150T 150T 146C 94A 146C 151T 146C 152C 263G 114T 263G 263G 195C 263G 146C 146C 146C 146C 146C 146C 146C 146C 146C 146C 151T 152C 152C 263G 263G 263G 150T 263G 152Y 146C 195C 195C 152C

195C 263G 195C 309.1C 309.1C 199C 199C 199C 295T 263G 263G 263G 188G 189G 228A 188G 228A 185A 228A 228A 228A 185A 228A 263G 228A 228A 152C 152C 152C 152C 195C 263G 195C 152C 195C 263G 309.1C 263G 309.1C 315.1C 263G 315.1C 152C 195C 152C 152C 152C 152C 152C 152C 152C 152C 152C 199C 199C 315.1C 315.1C 309.1C 263G 309.1C 263G 195C 263G 263G 263G

263G 309.1C 263G 315.1C 315.1C 204C 204C 263G 315.1C 295T 295T 295T 228A 228A 263G 228A 263G 228A 263G 263G 263G 188G 263G 295T 263G 263G 195C 195C 195C 263G 263G 309.1C 263G 263G 263G 295A 315.1C 309.1C 315.1C 497T 315.1C 497T 263G 263G 263G 263G 263G 263G 263G 263G 263G 263G 199C 204C 204C

309.1C 309.2C 315.1C

315.1C 315.1C

250C 207A 295T 462T 315.1C 315.1C 309.1C 263G 263G 295T 263G 295T 263G 295T 295T 295T 228A 295T 315.1C 295T 295T 215G 215G 215G 295T 309.1C 315.1C 309.1C 309.1C 309.1C 309.1C 497T 315.1C 497T

263G 239C 315.1C 489C 462T 462T 315.1C 295T 295T 309.1C 295T 315.1C 295T 309.1C 315.1C 309.1C 263G 315.1C 462T 315.1C 315.1C 263G 263G 263G 309.1C 315.1C

309.1C 250C 462T

315.1C 263G 489C

489C 489C 462T 309.1C 309.1C 315.1C 315.1C 462T 315.1C 315.1C 462T 315.1C 295T 462T 482C 462T 462T 295T 295T 295T 315.1C 524.1A

489C 315.1C 315.1C 462T 462T 489C 462T 462T 489C 462T 309.1C 482C 489C 482C 482C 310.1T 310.1T 310.1T 489C 524.2C

315.1C 315.1C 315.1C 315.1C

524.1A 524.1A 524.1A 497T

497T 524.1A

524.2C

315.1C 315.1C

524.1A

498del 315.1C

498del

309.1C 309.1C 263G

315.1C 315.1C 315.1C

315.1C 309.1C 309.2C 309.1C 198T 309.1C 315.1C 309.1C

385G 315.1C 315.1C 315.1C 263G 315.1C

574C 309.1C

576C 315.1C

462T 462T 489C 489C

489C 489C 524.1-4ACAC

489C 489C 489C 315.1C 489C 489C 489C 315.1C 315.1C 315.1C 523del

524.2C 524.2C 524.2C

497T 309.1C 309.1C 315.1C 315.1C 315.1C 315.1C 309.1C 315.1C 315.1C 315.1C 263G 263G 207A

315.1C

524.2C

524.1A 573.1-4CCCC 573.1C

524.1-4ACAC

573.1-5CCCCC

523del

524del

524.1-4ACAC

462T

489C

319C 319C 319C 524del

489C 489C 489C

513A 513A 513A

524.2C 573.2C

315.1C

169

f3G1 m2G4 m2H2 m3C4 m3D1 f2A6 m2B4 f2D4 f3G6 m1G1 m1H6 m2B5 f1B1 f1B2 f1E6 f3H4 m1E4 m1E6 m2E5 m2F2 m2G3 m3D4 f1H1 f3H5 m1B1 m3B3 f2G3 m1G6 m2H4 m3H2 f2D6 m2A6 m2E3 m3E3 m3F2 f1A1 f1A4 f1F3 f2G5 f3F4 m3C6 m3H3 f1H4 f2B2 f2H3 f3B2 f3E6 f3H3 m1G3 m1H2 m2F5 m3B5 f1B4 f3C5 m3B4 f2F2 f3C1 f3E2 f3F2 f3G5 m1B5 m1F6 f1D5 f1G5 f2C5

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

HV0 HV0 HV0 HV0 HV0 I I2 J1 J1 J1 J1 J1 J1c J1c J1c J1c J1c J1c J1c J1c J1c J1c J1c1 J1c1 J1c1 J1c1 J2a J2a J2a J2b K K K K K K1a K1a K1a K1a K1a K1a K1a K1c K1c K1c K1c K1c K1c K1c K1c K1c K1c K2b N1a N1a T T T T T T T T1 T1 T1

170

Table 2 (Continued ) Sample

HVS-I (16024–16569)

T1 T1 T1 T1 T1a T1a T1a T1a T1a T1a T2 T2 T2 T2 T2 T2 T2b T2b T2b T2b T2b T2b T2b T2b T2d T2d U U U U1a U1a U2e U2e U2e U2e U2e U2e U3a U3a U3a U4 U4 U4 U4 U4 U4 U4 U4 U4 U4 U4 U4 U4 U5a U5a U5a U5a1 U5a1 U5a1 U5a1 U5a1 U5a1

f3D2 f3H2 m1G2 m3B2 f1F4 f2H2 m1E1 m1H3 m2D6 m3C1 f1D6 f1H6 f2B3 f2F6 m2A5 m3H6 f2A5 f2G4 f2H6 f3F5 m3A6 m3E2 m3F4 m3F5 f3D3 m1B6 f2B6 f3B3 m2H3 f2E2 m1C1 f2H5 f3E3 m1D4 m2C3 m2C5 m2G6 m1C6 m2E1 m2F6 f1C4 f2B4 f2C2 f2E6 f2H1 f3H6 m1A6 m1B2 m1C4 m1D1 m1G4 m1H4 m3G2 f2F4 f3E5 m2E2 f1F6 f2C1 f2C3 f2G6 f3D4 f3F6

16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16126C 16111T 16126C 16126C 16519C 16519C 16192T 16126C 16126C 16051G 16051G 16051G 16051G 16051G 16051G 16093Y 16343G 16172C 16179T 16179T 16179T 16179T 16179T 16179T 16265G 16265G 16356C 16179T 16265G 16356C 16356C 16192T 16192T 16192T 16192T 16145A 16145A 16256T 16186T 16256T

16163G 16163G 16163G 16163G 16163G 16163G 16163G 16163G 16163G 16163G 16294T 16294T 16294T 16235G 16294T 16294T 16294T 16294T 16294T 16294T 16294T 16294T 16294T 16126C 16182C 16182C

16186T 16172C 16186T 16186T 16186T 16186T 16186T 16186T 16186T 16186T 16296T 16296T 16296T 16294T 16296T 16296T 16296T 16296T 16296T 16296T 16296T 16296T 16296T 16294T 16183C 16183C

HVS-II (1–576) 16189C 16186T 16189C 16189C 16189C 16189C 16189C 16189C 16189C 16189C 16519C 16519C 16519C 16296T 16324C 16519C 16304C 16304C 16304C 16304C 16304C 16304C 16304C 16296T 16189C 16189C

16291T 16182C 16182C 16129C 16129C 16129C 16129C 16129C 16129C 16173T 16390A 16343G 16266Y 16356C 16356C 16356C 16356C 16356C 16356C 16356C 16362C 16356C 16356C 16519C

16311C 16183C 16183C 16183C 16183C 16189C 16183C 16182C 16183C 16260T 16519C 16390A 16356C 16519C 16519C 16519C 16519C 16519C 16362C 16362C 16519C 16519C 16362C

16519C 16189C 16189C 16189C 16189C 16256T 16189C 16183C 16189C 16343G

16256T 16256T 16256T 16256T 16189C 16183C 16399G 16192T 16270T

16270T 16526A 16270T 16270T 16192T 16189C

16526A

16256T 16399G

16270T 16519C

16274A 16189C 16294T 16294T 16294T 16294T 16294T 16294T 16294T 16294T

16294T 16294T 16519C 16519C 16519C 16519C 16519C 16519C 16519C 16519C

16519C 16298C

16519C 16519C 16519C 16519C 16519C 16519C 16519C 16362C 16519C 16304C 16294T 16294T

16249C 16249C 16256T 16362C 16311C 16362C 16189C 16362C 16390A

16519C 16311C 16296T 16296T

16353T 16353T 16258C 16519C 16362C 16519C 16362C 16519C

16327T 16298C 16298C

16519C 16519C 16519C

16360T 16360T 16362C

16362C 16362C

16519C

16524G

16519C

16519C 16519C

16519C 16519C

16519C

16526A 16399G 16256T 16256T

16399G

16270T 16270T

16399G 16399G

16519Y

16291T

16399G

16519C

16519C

73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 44.1C 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G

263G 263G 152C 152C 152C 152C 152C 152C 152C 152C 150T 263G 263G 263G 263G 143A 263G 146C 195C 263G 204C 263G 73G 263G 195C 195C 152Y 195C 150T 183G 183G 152C 152C 152C 152C 152C 152C 150T 150T 150T 150T 150T 150T 150T 195C 195C 195C 195C 195C 150T 195C 195C 146C 263G 263G 263G 143A 195C 195C 263G 263G 263G

309.1C 309.1C 263G 234R 195C 195C 195C 195C 195C 195C 200G 309.1C 315.1C 309.1C 315.1C 263G 315.1C 263G 263G 315.1C 207A 309.1C 152C 309.1C 263G 263G 195C 263G 263G 263G 263G 183G 217C 263G 217C 217C 217C 199C 200G 263G 195C 195C 195C 195C 263G 263G 247A 247A 263G 195C 247A 263G 195C 309.1C 315.1C 315.1C 263G 263G 263G 309.1C 315.1C 315.1C

315.1C 315.1C 309.1C 263G 263G 263G 263G 263G 263G 263G 263G 315.1C

512G 315.1C 309.1C 309.1C 309.1C 309.1C 309.1C 309.1C 309.1C 315.1C

315.1C 315.1C 315.1C 315.1C 315.1C 309.2C 315.1C

315.1C 328G 315.1C

524.1A

524.2C

315.1C 309.1C

315.1C

263G 315.1C 263G 315.1C 315.1C 309.1C 263G 310C 315.1C 285T 285T 217C 263G 309.1C 263G 263G 263G 263G 263G 309.1C 263G 263G 263G 263G 309.1C 309.1C 263G 263G 309.1C 263G 263G 309.1C 228A 315.1C

315.1C

315.1C 315.1C

315.1C 310C 499A

499A

524.1A

315.1C 309.2C 309.1C 309.2C 315.1C 309.2C 309.2C 340T 523del 524.1A

315.1C 315.1C 315.1C 340T 315.1C 315.1C 508G 524del 524.2C

309.1C 309.1C 263G 309.1C 309.2C 309.1C 309.1C 315.1C 315.1C 315.1C 315.1C 309.1C 309.1C 315.1C 309.1C 315.1C 315.1C 315.1C 315.1C 315.1C 309.1C 315.1C 309.2C 263G

315.1C 315.1C 499A 315.1C 499A 499A 499A 499A 499A 315.1C 499A 315.1C 315.1C

315.1C 309.1C 309.1C 315.1C

309.2C 315.1C

315.1C 573.1–3CCC

524.1A

524.2C

499A 499A 524.1-4ACAC 499A

524.2C

340T 340T 508G 340T 340T 524.1A

508G 508G 524.1A 508G 492G 524.2C

524.1A 524.1-4ACAC

524.2C

524.1-4ACAC

524.1-4ACAC 524.1-4ACAC 499A 524.1-4ACAC 499A 499A

545A

524.1-4ACAC 524.1A

573.1C

524.2C

524.1A 524.1A 524.2C 524.1A 508G

524.2C 524.2C 524.2C

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

Hg

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

315.1C 263G

315.1C

315.1C 315.1C 263G 207A

263G

524.2C 263G 263G 207A 204C 315.1C 226C 315.1C 16278T 16278T 16255A 16278T 16274A 16223T 16223T 16223T

Variant positions from the rCRS are shown between 16024 and 16569 in HVS-I and 1 and 576 in HVS-II.

16519C 16300G

16292T 16519C 16519C 16519C 16291T

16519C 16519C 16278T 16519C

16519C 16519C 16519C

16270T 16278T 16270T 16270T 16270T 16270T 16295T 16362C 16223T 16519C 16223T 16189C 16189C 16189C

16519C

16278T 16519C 16278T 16278T

16325C

16399G 16519C 16519C 16519C 16398A 16398A

 The first two letters identify the geographical origin. For example, if population data from Vienna were generated, the first two letters would, e.g. be ‘‘Vi’’.  Following the city identifier, the plate identifier is added (1, 2, . . .).  Lastly, the 96-well plate position of the DNA extract is appended (e.g. ‘‘A1’’).

524.1A 207A 207A 204C 199C 309.2C 225A 263G 315.1C

315.1C 315.1C 315.1C 524.1A

524.1A 524.1A 524.1A 524.2C

524.2C 524.2C 524.2C

315.1C

sequencing primers that we used in former mtDNA population studies [14] turned out to provide insufficient sequence quality and were replaced (L16196, L00318, L00403 and H16410; Table 1(b)). To obtain a consistent and comprehensible nomenclature for all population samples generated for the EMPOP database, a sample naming procedure was conceived that takes the geographical origin and the position of the DNA extract in the 96-well DNA masterplate into account. This nomenclature enables an easy and straight-forward post-laboratory handling and organization of population data. Beginning with the creation of sample sheets for electrophoresis, where for new population data only the city and plate identifiers have to be replaced (which can automatically be done by e.g. MS Excel), also the further processing of the data (i.e. import into the sequence alignment software and the final import into the EMPOP database) can be organized consistently and unambiguously. So, for generating further population data, we recommend the following nomenclature to be applied:

315.1C 309.1C 309.1C 309.1C 315.1C 315.1C 315.1C 204C 204C 195C 195C 309.1C 195C 227G 309.2C

524.2C

315.1C 524.1A 315.1C 315.1C 263G 263G 263G 263G 309.1C 309.1C 309.1C 195C 195C 189G 194T 263G 189G 225A 309.1C

315.1C 309.1C

315.1C 315.1C 263G 315.1C 315.1C 309.1C 315.1C 263G 263G 189G 185A 185A 185A 263G 263G 263G 189G 189G 152C 189G 195C 153G 195C 263G 263G 263G 195C 263G 263G 263G 263G 150T 150T 150T 150T 150T 150T 150T 150T 150T 119C 150T 143A 152Y 153G 152C 153G 195C 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 73G 16399G 16320T 16399G 16399G 16293G 16311C 16270T 16270T

16399G 16270T 16270T 16256T 16256T 16399G 16399G 16270T 16270T

16270T 16256T 16256T 16249C 16249C 16270T 16270T 16192T 16192T 16270T 16209C 16270T 16220R 16189C 16192T 16189C 16292T 16292T 16192T 16292T 16189C 16183C 16183C 16183C 16256T 16192T 16192T 16192T 16192T 16256T 16256T 16189C 16189C 16261T 16189C 16189C 16189C 16183C 16189C 16129A 16223T 16223T 16093C 16223T 16092C 16182C 16182C 16134T f3G2 m1C5 m1D3 m1D6 m2G2 m2G5 m3C3 f1B5 f1D3 f2D2 f3A4 m2D5 m2H5 m3D5 m3D6 m3F3 f1F2 f2F1 m2C1 m2H1 m1C3 f1D2 m3D2 f1E5 U5a1 U5a1 U5a1 U5a1 U5a1 U5a1 U5a1 U5b U5b U5b U5b U5b U5b U5b U5b U5b W W W W X2 X2b X2c X2d

171

The sequencing reactions deriving from different sequencing primers are attached after the sample name and separated by an underscore. The full name of a sequence electropherogram obtained from e.g. sample Vi1A1 generated with the sequencing primer L15971 would thus be: ‘‘Vi1A1_L15971.ab1’’. The file extension ‘‘*.ab1’’ is automatically appended to file names by AB capillary electrophoresis instruments. 3.1. Polymorphisms in the entire mtDNA control region When taking dominant length variants in C-stretches into consideration, sequence comparisons led to the identification of 222 mitochondrial lineages as defined by 189 variable sites. On average, the samples showed 8.9 (95% CI 8.46–9.35) differences to the rCRS [15,16]. The mean pairwise difference between individuals was 9.32 (95% CI 9.28–9.36). Within the Austrian sequences, the ratio of sites manifesting transition mutations to those manifesting transversion mutations is 13.3. Point heteroplasmy was reproducibly detected in 18 out of 273 samples (6.6%)—f1C4: 16266Y; f1C6: 16278Y; f2B6: 152Y; f2C3: 16519Y; f2H4: 16172Y; f3F3: 16093Y; f3G1: 16311Y; f3G6: 16086Y and 16168Y; m1B3: 16355Y; m1B5: 152Y; m1C6: 16093Y; m2E5: 16232S; m2H1: 152Y: m2H5: 16220R; m3B2: 234R; m3B5: 16301Y; m3G6: 16216R; m3H3: 16362Y. We targeted the entire control region for this database to permit access to additional discriminatory variation that resides outside of HVS-I/HVS-II in the control region [31,36,37]. In the presented Austrian database, the entire control region discriminates 15 additional haplotypes compared to HVS-I/

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

172

HVS-II sequencing (206 vs 191 different haplotypes) disregarding cytosine insertions in the C-stretches (in HVS-I, HVS-II and HVS-III). The additional power of discrimination is not the only reason for targeting the mini-variable regions outside the control region, many phylogenetically informative sites also lie outside the commonly typed range 16024–16365 and 73–340. Haplogroup HV0 (former pre-V), for example, is characterized by the polymorphisms T16298C and T72C [38]. Many laboratories that amplify HVS-I and HVS-II separately start with the analysis of HVS-II at position 73, although position 72 has been sequenced and is clearly readable. Thus, when only HVS-II is typed from a sample sequence evaluation should start at least at position 72. Similar observations were made with other haplogroups: In this study, we confirm that U2e is among other diagnostic sites characterized by A508G, U3a by G16390A, U4 by G499A, U5a1 by A16399G, H1c by 477C, H5 by C456T, K1a by 497T, J by T489C, J1 by C462T and J1c1 by T482C (Table 2) [25,26]. Several years before, we [39] published an mtDNA population study on 101 Austrian Caucasians. In order to avoid redundancy we randomly selected from an independent set of samples. However, as identified during data analysis, we unintentionally included a sample in the new set that has also been typed in the first study (AUT69 from [39] is identical to m3G3). This is the mtDNA haplotype of a woman, who has accidentally been sampled twice within the past 5 years. Nevertheless, we decided to maintain this sample in the set of this study. 3.2. Random match probability (RMP) We have targeted the entire control region for this dataset to access additional discriminatory variation that resides outside of HVS-I/HVS-II [36,37]. The probability of a random match between two unrelated individuals from this Austrian dataset (n = 273) was calculated 1:89 for HVS-I + HVS-II and 1:125 for the entire CR (consistently disregarding C-insertions in HVS-I, HVS-II and HVS-III). The latter is comparable to RMPs computed for the entire control region databases from the Czech Republic (1:83; value from [30]) and Germany (1:96; [31]; RMP calculated from the data). The slight observed differences may confirm the assumption that the number of haplotypes increases with the sample size [40]. In addition, the observed decrease of discrimination power when only HVSI + HVS-II is analyzed was confirmed by the results from RMP

analysis of hypervariable regions I and II of the Czech (RMP = 1:67) and German (RMP = 1:66) population samples. The probability of a chance match (HVS-I + HVS-II) was calculated 1:59 for Bosnia-Herzegovina, 1:76 for Poland and 1:56 for Slovenia. 3.3. Comparison with other West-Eurasian populations In order to evaluate the diversity observed in Austria in relation to other West-Eurasian mtDNA control region variation, we compared the present Austrian sample set to databases of other West-Eurasian populations (Bosnia-Herzegovina, Czech Republic, Germany, Poland, and Slovenia). Only HVS-I and HVS-II data were considered, since not all databases included entire control region sequences. Pairwise comparisons showed a considerable number of matches between Austria and other European populations: 121 individuals (54 haplotypes) from Austria were also found in the databases from BosniaHerzegovina, the Czech Republic, Germany, Poland, and Slovenia. In particular, the most common haplotype in the Austrian population sample (263G, 315.1C) was also the most frequent profile in the other European populations (Austria: 7.7%, Bosnia-Herzegovina: 8.3%, Czech Republic: 8.6%, Germany: 9.5%, Poland: 9.0%, Slovenia: 7.8% of the population sample, respectively). On the other hand, the majority of sequences in the particular databases have not been observed in other databases (Austria: 55.7%, Bosnia-Herzegovina: 55.1%, Czech Republic: 76.3%, Germany: 57.5%, Poland: 62.3%, Slovenia: 52.4% of the population sample, respectively). Pairwise differences between and within populations were calculated with ARLEQUIN (Table 3). Whereas the populations from Austria, Bosnia-Herzegovina, Germany, Poland and Slovenia show on average 7.64 pairwise differences within their populations and 7.66 pairwise differences between the populations, the population sample from the Czech Republic shows 8.89 pairwise differences within its population and 8.29 differences to the other European populations. AMOVA was used to test for significant variation in the mtDNA distributions among the various populations (Table 4(a)). 99.7% of the variance observed among the six populations is attributable to differences within populations, and 0.3% ( pa < 0.05) represents differences among populations. The comparison of FST values from pairs of population samples (Table 4(b) and (c)) revealed that the Austrian

Table 3 Population average pairwise differences (16024–16368 and 71–340)

Austria Bosnia-Herzegovina Czech Republic Germany Poland Slovenia

Austria

Bosnia-Herzegovina

Czech Republic

Germany

Poland

Slovenia

7.81507 0.02895 0.00976 0.01856 0.03558 0.02732

7.59969 7.32642 0.03493 0.01334 0.02643 0.00699

8.36260 8.14344 8.89060 0.03489 0.01665 0.01971

7.84875 7.59919 8.40284 7.84530 0.01718 0.03010

7.78091 7.52743 8.29975 7.77762 7.67559 0.00742

7.69467 7.41603 8.22483 7.71257 7.60502 7.51963

Above diagonal: average number of pairwise differences between populations (PiXY); diagonal elements: average number of pairwise differences within population (PiX); below diagonal: corrected average pairwise difference (PiXY  (PiX + PiY)/2).

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

173

Table 4 AMOVA results Source of variation

d.f.

(a) Design and results (d.f. stands for degrees of freedom) Among populations 5 Within populations 1064 Total

1069

(b) Population pairwise FSTs Austria Bosnia-Herzegovina Czech Republic Germany Poland Slovenia

Sum of squares

Variance components

Percent of variation

29.289 4136.307

0.01136 Va 3.89116 Vb

0.29 99.71

4165.595

3.90253

Austria

Bosnia-Herzegovina

Czech Republic

Germany

Poland

Slovenia

0.00000 0.00369 0.00168 0.00237 0.00457 0.00340

0.00000 0.00482 0.00170 0.00344 0.00090

0.00000 0.00462 0.00258 0.00249

0.00000 0.00222 0.00378

0.00000 0.00091

0.00000

Austria

Bosnia-Herzegovina

Czech Republic

Germany

Poland

Slovenia

* 0.02441 0.17773 0.05176 0.00391 0.06445

* 0.02344 0.11328 0.02734 0.60645

* 0.02441 0.08789 0.14844

* 0.06348 0.04492

* 0.25391

*

a

(c) FST P values Austria Bosnia-Herzegovina Czech Republic Germany Poland Slovenia a

Significant FST P values are depicted in bold.

population sample displayed significant differences in mtDNA distributions compared to the population databases from Bosnia-Herzegovina and Poland. Additionally, the Bosnian database showed statistically significant differences to the population samples from Poland and the Czech Republic, which also displayed differences to the German database. Lastly, the German sequences also showed differences in population substructure to sequences from Slovenia. The slight differences in mtDNA substructure between the different populations might be explained by the relatively small sizes of the samples compared to the sizes of the populations they were drawn from. The small sample sizes might thus represent faintly different cutouts of the distribution of the mitochondrial haplogroups. However, the inter-population variability calculated for these European populations is in agreement with former estimates from other European populations [41] and is low compared to sub-Saharan African populations [42,13]. 3.4. Haplogroup assignment Much of the evolutionary change in the mtDNA pools of West-Eurasian populations during the past centuries has been introduced by lineage redistribution—both within populations, caused by drift, and between populations, caused by migration [43]. The basic phylogenetic structure of West-Eurasian mtDNA lineages has been revealed by a number of recent studies [44,45,18,46]. The analysis of 16 diagnostic mtDNA coding region SNPs allowed us to assign the major haplogroup status of the CR sequences [28], which were in agreement with the expected haplogroup distribution from West-Eurasian populations (H, I, J, K, N1a, T, U, V, W, and X) [46–50,28,23–25].

The most common haplogroup, cluster HV (including samples belonging to haplogroups HV0 (former pre-V), H* and sub-haplogroups of H), was observed in 45.4% of the Austrian sample set. Samples belonging to haplogroup H, which could not be resolved genealogically in subclades [23,24,51], were assigned to haplogroup H*. Haplogroups observed at intermediate levels included clusters U (19.8%), T (13.2%), J (8.4%), and K (8.4%). The haplogroups observed less frequently included I (0.7%), N1a (0.7%), W (1.5%), and X (1.5%). One sample (f1G6) that could not be unambiguously assigned by the 16 coding region SNPs was identified by control region polymorphisms to belong to the Asian haplogroup C (0.4%; Table 2). Two samples that were assigned to haplogroup I by the 16 coding region SNPs were identified to belong to haplogroup N1a by control region markers. This initial misclassification can be explained by the fact that by the time of the SNP study [28], the marker G1719A was used to classify haplogroups I and X [47,45,50]. According to a later study however [25], the marker G1719A turned out as a deeprooting marker defining the two branches N1 and N5, where haplogroup I is a only sub-lineage of haplogroup N1. A refined classification into nested sub-haplogroups could be reached for haplogroups I, J, K, T, U and V. Due to the lack of a detailed description of haplogroups W and X, samples belonging to these haplogroups could not be further subdivided. 3.5. Phylogenetic analyses The phylogenetic check for artificial recombination applied here provides a fast method for identifying samples containing DNA-stretches from two different persons. Such a quality check is especially valuable for mtDNA databases, where the

174

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175

two hypervariable regions have been amplified and sequenced separately. Our approach to sequence generation involves the amplification of the entire control region in one step and the generation of redundant sequence information by sequencing overlapping DNA fragments. Thus, the occurrence of artificial recombination of DNA stretches deriving from different persons can be excluded. However, a byproduct of this phylogenetic check for artificial recombination is that all samples are assigned with a mitochondrial haplogroup. This phylogenetic haplogroup determination procedure again is approved and completed by the results of the haplogroup determination algorithm based on propositional logic. The combination of the two programs compensates for the weakness of each single program: the phylogenetic inference fails if the background dataset does not contain a sequence of similar ethnic origin to the sample in question; the logic approach does not take the different mutation rates of the individual diagnostic positions into full account and might thus lead the user to the wrong tip of the mitochondrial tree. Thus, both approaches can facilitate the assignment of samples to mitochondrial haplogroups; however, the final decision needs to be based on a profound understanding of the human mitochondrial genome, its mutation rates and evolutionary patterns. Acknowledgements We thank Verena Lubei, Anna Ko¨nig and Dr. Cordula Eichmann (Institute of Legal Medicine, Innsbruck Medical University, Austria) for help with sequence evaluation. References [1] M.M. Holland, T.J. Parsons, Mitochondrial DNA sequences analysis— validation and use for forensic casework, Forensic Sci. Rev. 11 (1) (1999) 21–49. [2] H.-J. Bandelt, W. Parson, Fehlerquellen mitochondrialer DNA-Datensa¨tze und evaluation der mtDNA-Datenbank ‘‘D-Loop-BASE’’ [sources of errors in mitochondrial DNA datasets and evaluation of the mtDNA database ‘‘D-Loop-BASE’’], Rechtsmedizin 14 (2004) 251–255. [3] H.-J. Bandelt, A. Salas, S. Lutz-Bonengel, Artificial recombination in forensic mtDNA population databases, Int. J. Legal Med. 118 (5) (2004) 267–273. [4] Y.-G. Yao, C.M. Bravi, H.-J. Bandelt, A call for mtDNA data quality control in forensic science, Forensic Sci. Int. 141 (1) (2004) 1–6. [5] A. Carracedo, W. Ba¨r, P. Lincoln, W. Mayr, N. Morling, B. Olaisen, et al., DNA commission of the international society for forensic genetics: guidelines for mitochondrial DNA typing, Forensic Sci. Int. 110 (2) (2000) 79–85. [6] G. Tully, W. Ba¨r, B. Brinkmann, A. Carracedo, P. Gill, N. Morling, et al., Considerations by the European DNA profiling (EDNAP) group on the working practices, nomenclature and interpretation of mitochondrial DNA profiles, Forensic Sci. Int. 124 (1) (2001) 83–91. [7] W. Parson, A. Brandsta¨tter, A. Alonso, N. Brandt, B. Brinkmann, A. Carracedo, et al., The EDNAP mitochondrial DNA population database (EMPOP) collaborative exercises: organisation, results and perspectives, Forensic Sci. Int. 139 (2–3) (2004) 215–226. [8] A. Salas, L. Prieto, M. Montesino, C. Albarra´n, E. Arroyo, M.R. ParedesHerrera, et al., Mitochondrial DNA error prophylaxis: assessing the causes of errors in the GEP’02-03 proficiency testing trial, Forensic Sci. Int. 148 (2–3) (2005) 191–198.

[9] W. Parson, A. Brandsta¨tter, M. Pircher, M. Steinlechner, R. Scheithauer, EMPOP—the EDNAP mtDNA population database concept for a new generation, high-quality mtDNA database, Int. Congr. Ser. 1261 (2004) 106–108. [10] E. Ambach, W. Parson, H. Niedersta¨tter, B. Budowle, Austrian Caucasian population data for the quadruplex plus amelogenin: refined mutation rate for HumvWFA31/A, J. Forensic Sci. 42 (6) (1997) 1136–1139. [11] W.J. Jones, J.A. Irwin-Ross, F.A. Love, A.B. Welsh, M.M. Holland, T.J. Parsons, Development of an efficient, high-throughput strategy for sequence analysis of the entire human mitochondrial DNA control region. in: Poster, 10th International Symposium on Human Identification, Promega Corp., Orlando, FL, September 27–October 1, 1999. [12] E.T. Richon, N. Abassi, A. Coute, T.P. McMahon, J.A. Irwin, S.M. Barritt, et al., Validation of the Tecan Genesis robotic sample processor for automated cycle sequencing of mtDNA database samples, in: Poster, 14th International Symposium on Human Identification, Promega Corp., Phoenix, AZ, September 28–October 3, 2003. [13] A. Brandsta¨tter, C.T. Peterson, J.A. Irwin, S. Mpoke, D.K. Koech, W. Parson, T.J. Parsons, Mitochondrial DNA control region sequences from Nairobi (Kenya): inferring phylogenetic parameters for the establishment of a forensic database, Int. J. Legal Med. 118 (5) (2004) 294–306. [14] A. Brandsta¨tter, H. Niedersta¨tter, W. Parson, Monitoring the inheritance of heteroplasmy by computer-assisted detection of mixed basecalls in the entire human mitochondrial DNA control region, Int. J. Legal Med. 118 (1) (2004) 47–54. [15] S. Anderson, A.T. Bankier, B.G. Barrell, M.H. de Bruijn, A.R. Coulson, J. Drouin, et al., Sequence and organization of the human mitochondrial genome, Nature 290 (5806) (1981) 457–465. [16] R.M. Andrews, I. Kubacka, P.F. Chinnery, R.N. Lightowlers, D.M. Turnbull, N. Howell, Reanalysis and revision of the Cambridge reference sequence for human mitochondrial DNA, Nat. Genet. 23 (2) (1999) 147. [17] M. Steinlechner, W. Parson, Automation and high through-put for a DNA database laboratory: development of a laboratory information management system, Croat. Med. J. 42 (3) (2001) 252–255. [18] A. Helgason, a.S. Sigureth, J.R. Gulcher, R. Ward, K. Stefansson, mtDNA and the origin of the Icelanders: deciphering signals of recent population history, Am. J. Hum. Genet. 66 (3) (2000) 999–1016. [19] B.A. Malyarchuk, T. Grzybowski, M.V. Derenko, J. Czarny, M. Wozniak, D. Miscicka-Sliwka, Mitochondrial DNA variability in Poles and Russians, Ann. Hum. Genet. 66 (Pt. 4) (2002) 261–283. [20] M.V. Derenko, T. Grzybowski, B.A. Malyarchuk, I.K. Dambueva, G.A. Denisova, J. Czarny, et al., Diversity of mitochondrial DNA lineages in South Siberia, Ann. Hum. Genet. 67 (5) (2003) 391–411. [21] M.V. Derenko, B.A. Malyarchuk, I.K. Dambueva, I.A. Zakharov, Structure and diversity of the mitochondrial gene pools of south Siberians, Dokl. Biol. Sci. 393 (2003) 557–561. [22] B.A. Malyarchuk, T. Grzybowski, M.V. Derenko, J. Czarny, K. Drobnic, D. Miscicka-Sliwka, Mitochondrial DNA variability in Bosnians and Slovenians, Ann. Hum. Genet. 67 (5) (2003) 412–425. [23] A. Achilli, C. Rengo, C. Magri, V. Battaglia, A. Olivieri, R. Scozzari, et al., The molecular dissection of mtDNA haplogroup H confirms that the Franco-Cantabrian glacial refuge was a major source for the European gene pool, Am. J. Hum. Genet. 75 (5) (2004) 910–918. [24] E.-L. Loogva¨li, U. Roostalu, B.A. Malyarchuk, M.V. Derenko, T. Kivisild, E. Metspalu, et al., Disuniting uniformity: a pied cladistic canvas of mtDNA haplogroup H in Eurasia, Mol. Biol. Evol. 21 (11) (2004) 2012– 2021. [25] M.G. Palanichamy, C. Sun, S. Agrawal, H.-J. Bandelt, Q.-P. Kong, F. Khan, et al., Phylogeny of mitochondrial DNA macrohaplogroup N in India, based on complete sequencing: implications for the peopling of South Asia, Am. J. Hum. Genet. 75 (6) (2004) 966–978. [26] A. Achilli, C. Rengo, V. Battaglia, M. Pala, A. Olivieri, S. Fornarino, et al., Saami and berbers—an unexpected mitochondrial DNA link, Am. J. Hum. Genet. 76 (5) (2005) 883–886. [27] H.-J. Bandelt, A. Achilli, Q.-P. Kong, A. Salas, S. Lutz-Bonengel, C. Sun, et al., Low ‘‘penetrance’’ of phylogenetic knowledge in mitochondrial disease studies, Biochem. Biophys. Res. Commun. 333 (1) (2005) 122– 130.

A. Brandsta¨tter et al. / Forensic Science International 166 (2007) 164–175 [28] A. Brandsta¨tter, T.J. Parsons, W. Parson, Rapid screening of mtDNA coding region SNPs for the identification of west European Caucasian haplogroups, Int. J. Legal Med. 117 (5) (2003) 291–298. [29] M. Stoneking, D. Hedgecock, R.G. Higuchi, L. Vigilant, H.A. Erlich, Population variation of human mtDNA control region sequences detected by enzymatic amplification and sequence-specific oligonucleotide probes, Am. J. Hum. Genet. 48 (2) (1991) 370–382. [30] T. Vanecek, F. Vorel, M. Sip, Mitochondrial DNA D-loop hypervariable regions: Czech population data, Int. J. Legal Med. 118 (1) (2004) 14–18. [31] S. Lutz, H.-J. Weisser, J. Heizmann, S. Pollak, Location and frequency of polymorphic positions in the mtDNA control region of individuals from Germany, Int. J. Legal Med. 111 (2) (1998) 67–77 (Errata in Int. J. Legal Med. 111 (1998) 286 and Int. J. Legal Med. 112 (1999) 145–150). [32] S. Schneider, D. Roessli, L. Excoffier, Arlequin ver. 2.0: a software for population genetic data analysis. Genetics and Biometry Laboratory, University of Geneva, Switzerland, 2000. [33] H.-J. Bandelt, A. Salas, C. Bravi, Problems in FBI mtDNA database, Science 305 (5689) (2004) 1402–1404. [34] P.J. Waddell, M.A. Steel, General time-reversible distances with unequal rates across sites: mixing gamma and inverse Gaussian distributions with invariant sites, Mol. Phylogenet. Evol. 8 (3) (1997) 398–414. [35] H.-J. Bandelt, P. Lahermo, M. Richards, V. Macaulay, Detecting errors in mtDNA data by phylogenetic analysis, Int. J. Legal Med. 115 (2) (2001) 64–69. [36] S. Lutz, H.-J. Weisser, J. Heizmann, S. Pollak, A third hypervariable region in the human mitochondrial D-loop, Hum. Genet. 101 (3) (1997) 384. [37] M.D. Coble, R.S. Just, J.E. O’Callaghan, I.H. Letmanyi, C.T. Peterson, J.A. Irwin, T.J. Parsons, Single nucleotide polymorphisms over the entire mtDNA genome that increase the power of forensic testing in Caucasians, Int. J. Legal Med. 118 (3) (2004) 137–146. [38] A. Torroni, H.-J. Bandelt, V. Macaulay, M. Richards, F. Cruciani, C. Rengo, et al., A signal, from human mtDNA, of postglacial recolonization in Europe, Am. J. Hum. Genet. 69 (4) (2001) 844–852. [39] W. Parson, T.J. Parsons, R. Scheithauer, M.M. Holland, Population data for 101 Austrian Caucasian mitochondrial DNA d-loop sequences: application of mtDNA sequence analysis to a forensic case, Int. J. Legal Med. 111 (3) (1998) 124–132.

175

[40] L. Pereira, C. Cunha, A. Amorim, Predicting sampling saturation of mtDNA haplotypes: an application to an enlarged Portuguese database, Int. J. Legal Med. 118 (3) (2004) 132–136. [41] T. Melton, M. Wilson, M. Batzer, M. Stoneking, Extent of heterogeneity in mitochondrial DNA of European populations, J. Forensic Sci. 42 (3) (1997) 437–446. [42] T. Melton, C. Ginther, G. Sensabaugh, H. Soodyall, M. Stoneking, Extent of heterogeneity in mitochondrial DNA of sub-Saharan African populations, J. Forensic Sci. 42 (4) (1997) 582–592. [43] A. Helgason, E. Hickey, S. Goodacre, V. Bosnes, K. Stefansson, R. Ward, B. Sykes, mtDna and the islands of the North Atlantic: estimating the proportions of Norse and Gaelic ancestry, Am. J. Hum. Genet. 68 (3) (2001) 723–737. [44] M. Richards, V. Macaulay, H.-J. Bandelt, B.C. Sykes, Phylogeography of mitochondrial DNA in western Europe, Ann. Hum. Genet. 62 (3) (1998) 241–260. [45] V. Macaulay, M. Richards, E. Hickey, E. Vega, F. Cruciani, V. Guida, et al., The emerging tree of West Eurasian mtDNAs: a synthesis of controlregion sequences and RFLPs, Am. J. Hum. Genet. 64 (1) (1999) 232–249. [46] S. Finnila¨, M.S. Lehtonen, K. Majamaa, Phylogenetic network for European mtDNA, Am. J. Hum. Genet. 68 (6) (2001) 1475–1484. [47] A. Torroni, K. Huoponen, P. Francalacci, M. Petrozzi, L. Morelli, R. Scozzari, et al., Classification of European mtDNAs from an analysis of three European populations, Genetics 144 (4) (1996) 1835–1850. [48] D.C. Wallace, M.D. Brown, M.T. Lott, Mitochondrial DNA variation in human evolution and disease, Gene 238 (1) (1999) 211–230. [49] M.W. Allard, K. Miller, M. Wilson, K.L. Monson, B. Budowle, Characterization of the Caucasian haplogroups present in the SWGDAM forensic mtDNA dataset for 1771 human control region sequences. Scientific working group on DNA analysis methods, J. Forensic Sci. 47 (6) (2002) 1215–1223. [50] C. Herrnstadt, J.L. Elson, E. Fahy, G. Preston, D.M. Turnbull, C. Anderson, et al., Reduced-median-network analysis of complete mitochondrial DNA coding-region sequences for the major African, Asian, and European haplogroups, Am. J. Hum. Genet. 70 (5) (2002) 1152–1171. [51] K. Tambets, S. Rootsi, T. Kivisild, H. Help, P. Serk, E.-L. Loogva¨li, et al., The western and eastern roots of the Saami–the story of genetic ‘‘outliers’’ told by mitochondrial DNA and Y chromosomes, Am. J. Hum. Genet. 74 (4) (2004) 661–682.

Forensic Science International 166 (2007) 176–181 www.elsevier.com/locate/forsciint

Diversity of 26-locus Y-STR haplotypes in a Nepalese population sample: Isolation and drift in the Himalayas Emma J. Parkin a, Thirsa Kraayenbrink b, Jean Robert M.L. Opgenort c, George L. van Driem c, Nirmal Man Tuladhar d, Peter de Knijff b, Mark A. Jobling a,* a

b

Department of Genetics, University of Leicester, University Road, Leicester LE1 7RH, UK MGC-Department of Human and Clinical Genetics, Leiden University Medical Centre, The Netherlands c Himalayan Languages Project, Leiden University, The Netherlands d Centre for Nepal and Asian Studies (CNAS) of Tribhuvan University (TU), Kirtipur, Nepal Received 30 March 2006; received in revised form 8 May 2006; accepted 9 May 2006 Available online 15 June 2006

Abstract Twenty-six Y-chromosomal short tandem repeat (STR) loci were amplified in a sample of 769 unrelated males from Nepal, using two multiplex polymerase chain reaction (PCR) assays. The 26 loci gave a discriminating power of 0.997, with 59% unique haplotypes, and the highest frequency haplotype occurring 12 times. We identified novel alleles at four loci, microvariants at a further two, and nine examples of amelogenin-Y deletions (1.2%). Comparison with a similarly sized Bhutanese sample typed with the same markers suggested histories of isolation and drift, with drift having a greater effect in Bhutan. Extended (11-locus) haplotypes for the Nepalese samples have been submitted to the Y-STR Haplotype Reference Database (YHRD). # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Y chromosome; STRs; Microsatellites; Haplotype; Nepal; Bhutan; Himalayas

1. Introduction The analysis of multiple Y-chromosomal short tandem repeats (STRs) provides informative male-specific DNA profiles in forensic analysis. As well as possessing high discriminating power in distinguishing individuals, haplotypes defined by STRs can provide information about likely geographical origin, since they are often concentrated in particular populations or regions. Population databases of Y haplotypes [1] are increasing in size and coverage, greatly contributing to the utility of Ychromosomal analysis in forensic casework. In this study we describe alleles at 26 Y-STRs, and properties of the haplotypes they define, in a large sample of a previously unrepresented population, that of Nepal in the Himalayas. Eleven-locus haplotypes have been submitted to the Y-STR Haplotype

* Corresponding author. Tel.: +44 116 252 3427; fax: +44 116 252 3378. E-mail address: [email protected] (M.A. Jobling). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.007

Reference Database (YHRD), and full data are available from the authors on request. Our report follows guidelines for the publication of population data [2]. Sampling and Y-chromosomal analysis of 769 Nepalese males was undertaken as part of a larger collaborative project [3] investigating genetic diversity in Himalayan populations within the framework of their cultural and linguistic diversity [4]. Here we describe our initial findings with Y-STRs, treating the Nepalese sample as a single population; future publications will explore genetic relationships between subpopulations of the Himalayas. The sample represents 15 distinct ethnolinguistic groups widely distributed throughout Nepal, with 75% of sampled individuals speaking languages belonging to the Tibeto-Burman family, and the remainder speaking IndoEuropean languages. In this study, we employ the same set of Y-STRs as that used recently to analyse 856 Bhutanese males [5]. This allows us to carry out a preliminary comparison of diversity and haplotype sharing between these two Himalayan samples.

Table 1 Frequencies of alleles at 22 of the 26 Y-STRs Allele

ha h(Bh)b a b

388

390

391

392

393

425

426

434

435

436

437

0.022

0.001 0.044 0.644 0.241 0.066 0.004

0.053 0.776 0.164 0.005 0.001

0.563 0.004 0.289 0.068 0.020 0.046 0.008 0.001 0.003

0.072 0.261 0.059 0.023 0.497 0.049 0.016 0.001

0.009 0.675 0.211 0.100 0.005

0.001 0.948 0.018

0.003 0.826 0.170 0.001

0.087 0.870 0.016 0.027

0.001 0.982 0.017

0.001 0.978 0.021

0.003

438 0.001 0.003 0.085 0.168 0.719 0.025

0.001 0.386 0.544 0.065 0.001

439

447

448

0.004 0.001 0.511 0.270 0.176 0.038

0.151 0.248 0.390 0.192 0.016 0.003 0.003 0.003 0.005

0.005 0.051 0.490 0.321 0.124 0.008 0.001

460

0.057 0.501 0.144 0.086 0.100 0.072 0.010 0.004 0.001 0.013 0.001

461

462

389I

0.018 0.008 0.140 0.645 0.160 0.027

0.001 0.217 0.684 0.096 0.001

0.003 0.014 0.524 0.290 0.165 0.004

0.003 0.046 0.025 0.242 0.614 0.064 0.003

389II-I

H4.1

0.001 0.168 0.550 0.189 0.087 0.005

0.022 0.280 0.586 0.104 0.008

0.001 0.004 0.005 0.005

E.J. Parkin et al. / Forensic Science International 166 (2007) 176–181

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 21.4 22.4 23.4 9–12 10–11 11–12 12–13 13–14 20–21 Null

19

0.001 0.001 0.004

0.017 0.521 0.604

0.593 0.518

0.639 0.569

0.368 0.421

0.673 0.546

0.490 0.442

0.040 0.187

0.289 0.244

0.235 0.244

0.036 0.046

0.043 0.092

0.551 0.553

0.447 0.452

0.726 0.713

0.702 0.663

0.553 0.590

0.633 0.679

0.536 0.434

0.476 0.363

0.614 0.592

0.626 0.504

0.567 0.598

Calculation of gene diversity, h, excludes null alleles and duplications. Comparative Bhutanese values from [5].

177

178

E.J. Parkin et al. / Forensic Science International 166 (2007) 176–181

2. Materials and methods

Table 2 Frequencies of genotypes at DYS385 and YCAII

2.1. DNA samples

Genotype

DYS385

10–14 11–11 11–12 11–13 11–14 11–15 11–16 11–18 11–19 11–20 12–12 12–13 12–14 12–15 12–16 12–17 12–18 12–19 12–20 13–13 13–14 13–15 13–16 13–17 13–18 13–19 13–20 13–21 13–22 13–23 14–14 14–15 14–16 14–17 14–18 14–19 14–20 14–22 15–15 15–16 15–17 15–18 15–19 15–20 15–21 16–16 16–17 16–18 16–19 16–20 16–22 17–17 17–18 17–19 17–20 17–21 17–22 17–23 18–18 18–19 18–20 19–19 20–20 20–21

0.005 0.017 0.007 0.004 0.070 0.001 0.001 0.003 0.001 0.001 0.004 0.003 0.009 0.001 0.027 0.029 0.014 0.007 0.012 0.022 0.022 0.003 0.027 0.056 0.182 0.177 0.060 0.014 0.001 0.003 0.004 0.007 0.010 0.013 0.044 0.031 0.027 0.001 0.004 0.009 0.007 0.013 0.003 0.007 0.001 0.001 0.004 0.004 0.003 0.003

Seven hundred and sixty-nine Bhutanese males provided blood samples with informed consent, and DNA was extracted as described [3]. DNA samples from collections of the authors, including Y Chromosome Consortium (YCC) cell lines [6], were used as haplotype reference materials. 2.2. Y-STR multiplexes Two PCR multiplexes (a 20plex [7] and a partially overlapping 14plex [5]) were used to type 26 Y-STRs, as follows: DYS19, DYS385a/b, DYS388, DYS389I, DYS389II, DYS390, DYS391, DYS392, DYS393, DYS425, DYS426, DYS434, DYS435, DYS436, DYS437, DYS438, DYS439, DYS447, DYS448, DYS460, DYS461, DYS462, YCAIIa/b, and Y-GATA-H4.1. The eleven Y-STR markers in the European ‘extended haplotype’ (http://www.yhrd.org/) are indicated in bold. The 14plex includes the amelogenin sex test. Full details of the protocol are given by Parkin et al. [5]. 2.3. Y-STR nomenclature Allele nomenclature (explained fully in Parkin et al. [5]) was according to Butler et al. [7] and Bosch et al. [8], with the exception of DYS439, DYS448 and Y-GATA-H4.1, where nomenclature was changed for compatibility with ISFG recommendations [9]. Compared to Butler et al. [7], seven repeats were subtracted from DYS439, three subtracted from DYS448, and eight added to Y-GATA-H4.1. 2.4. Calculations Gene diversity and haplotype diversity were calculated using Arlequin [10]. A median-joining network was constructed using Network 4.0 ([11] http://www.fluxus-engineering.com/sharenet.htm), and the weighting scheme described by Qamar et al. [12]. 3. Results and discussion 3.1. Diversity of alleles Tables 1 and 2 show the allele frequency distributions for all the Y-STRs studied. Diversities of individual STRs are comparable with those observed in a recently studied Bhutanese sample: DYS385 (when considered as a genotype, Table 2) is the most diverse marker within the Y-STR set, with a gene diversity (h) of 0.915, and the most polymorphic singlelocus marker is DYS439 (h = 0.726). Previously unreported alleles (defined with reference to Butler [13], Parkin et al. [5] and STRBase, http://www.cstl.nist.gov/biotech/strbase/index.htm) were found at four loci, as follows: DYS426 (allele 13), DYS437 (allele 11), DYS439 (allele 15), DYS447 (alleles 17, 18 and 19).

0.003

0.001 0.001

YCAII

0.009

0.014

0.008 0.003 0.010 0.001 0.036 0.029 0.606 0.113 0.091 0.001 0.001 0.012 0.025 0.001 0.036 0.001 0.001

E.J. Parkin et al. / Forensic Science International 166 (2007) 176–181 Table 2 (Continued ) Genotype

DYS385

13–17.2 13–18.2

0.001 0.016

h h(Bh)a

0.915 0.921

a

YCAII

0.607 0.524

Comparative Bhutanese values from [5].

‘Null’ alleles or multiple peaks were reproducibly obtained at a number of loci. For DYS448, three individuals carried null alleles, while one carried both alleles 20 and 21. For DYS461, one individual carried both alleles 13 and 14. As observed previously [5], DYS425 exhibits a relatively high frequency of various nulls and duplications. Microvariants (partial alleles) were observed at two loci (Tables 1 and 2) and confirmed in uniplex assays after initial detection in multiplexes. Those at DYS385 were not investigated further, but those at DYS447 were analysed by sequencing, and shown to result from a deletion of 1 bp within the pentanucleotide repeat array [5]. 3.2. AMELY deletion chromosomes Nine chromosomes showed absence of the amelogenin Y (AMELY) peak in electropherograms. Analysis of sequencetagged sites revealed that these chromosomes carry interstitial

179

deletions of Yp including the AMELY locus (data not shown); none showed null Y-STR alleles, however, which is consistent with the size and location of known AMELY deletions with respect to the position of Y-STR loci [14]. A previous study has found AMELY deletions at a frequency of 2% in India [15], so our finding of deletions at 1.2% frequency in Nepal is not unexpected; in contrast, however, none were found in our previous study of Bhutan [5]. These AMELY deletion chromosomes form part of a large set that is currently being characterised, and will be described fully elsewhere. 3.3. Diversity of haplotypes Haplotype diversity (equivalent to power of discrimination, PD) was calculated, omitting chromosomes carrying null alleles and duplications. This provided a sample size of 741. For the full set of 26 Y-STRs, there are 437 unique haplotypes (59.0%), and PD is 0.9970. The corresponding values for the 20plex [7], extended (11-locus) haplotype and minimal (9locus) haplotype are shown in Fig. 1. Fig. 1 also shows the distribution of haplotypes present more than once in the dataset. Despite the large number of loci used here, in the 741 males one 26-locus haplotype is shared by 12 individuals (Fig. 1a), and a further 13 haplotypes are shared by between 5 and 9 individuals; notably, all these common haplotypes are restricted to particular subpopulations, illustrating the influence of drift. Reduction to 11-locus extended haplotypes allows a global search within the YHRD (release 18): this fails to find matches for three of the six most common Nepalese extended haplotypes (frequency 10), consistent with isolation and drift. 3.4. Comparison of Y-STR datasets on Nepal and Bhutan The availability of large Y-STR haplotype datasets on Nepalese and Bhutanese samples allows us to make

Fig. 1. Haplotype diversity for (a) all 26 STRs and the 20plex, and (b) the extended and minimal haplotypes. Histograms show the frequency distributions of haplotypes present more than once in the dataset.

Fig. 2. Percentage of unique haplotypes in Nepal compared to Bhutan. Haplotypes containing null or duplicated alleles are omitted, giving total sample sizes of 741 and 802, respectively. The error bars represent plus or minus one binomial standard error.

180

E.J. Parkin et al. / Forensic Science International 166 (2007) 176–181

Fig. 3. Median-joining network of haplotypes shared between Nepal and Bhutan. Note that the 40 shared extended haplotypes described in the text are reduced to 28 when the bilocal Y-STR DYS385 is removed for network construction. Circles represent Y-STR haplotypes (based on DYS19, DYS389I, DYS389II-I, DYS390, DYS391, DYS392, DYS393, DYS438, DYS439), with area proportional to number of instances. Lines represent Y-STR mutational steps.

comparisons between the frequencies and distributions of alleles and haplotypes in these two Himalayan populations. Allele distributions at individual loci are similar between the Nepalese and Bhutanese samples, but this gives little information about population relationships. Particular rare and distinctive alleles may carry more information, because they probably reflect identity-by-descent: a good example of this is the sharing of microvariants at DYS447 [5], but apart from this there is little evidence for specific inter-population sharing. Comparison of haplotype distributions reveals a striking difference between the two populations. The proportion of unique haplotypes in the Nepalese sample is significantly greater than that in the Bhutanese, for all four haplotype resolutions considered (Fig. 2). For example, for the extended haplotype there are 41.8% (1.8%) unique haplotypes in Nepal, but only 23.3% (1.5%) in Bhutan. This is explained by the presence of several common haplotypes at high frequency in Bhutan: in the Nepalese dataset, the most common extended haplotypes are each present in 13 individuals, while in the Bhutanese there are haplotypes present in 15, 16, 24 (two instances) and 27 individuals [5]. There are no 26-locus haplotypes shared between Nepal and Bhutan, indicating an absence of very recent gene flow. However, forty extended haplotypes are shared between the two samples, and their relationships (omitting the bilocal marker DYS385) are illustrated in a median-joining network in Fig. 3. Most of them fall into one large cluster, with haplotypes linked by single mutational steps, probably representing a common YSNP haplogroup. Other shared haplotypes are more widely spread, and may represent several different haplogroups. To ask if these shared extended haplotypes are more generally common and widespread, we sought matches for the six most predominant examples (combined frequency >10) within the YHRD. Three of the six haplotypes find a total of six

exact matches, all within populations originating from China or the Indian subcontinent. We also find a total of 30 one-step mutational neighbours for five of the six haplotypes, all of Asian origin. One haplotype finds neither exact matches nor one-step neighbours. Thus, the common haplotypes shared between Nepal and Bhutan are Asian-specific, but not generally frequent. 3.5. Concluding remarks Our study emphasises the discriminating power of highresolution Y-STR typing, and provides the first substantial dataset on a Nepalese sample. The comparison of Nepalese and Bhutanese datasets reveals an interesting overall picture of isolation and drift within these Himalayan populations, with drift having a greater effect in Bhutan than Nepal. Haplotype sharing provides evidence of some gene flow between Nepal and Bhutan, or possibly of gene flow into both from some other population. Further light will be thrown on these relationships when Y-SNP data become available. Acknowledgements We thank all DNA donors, and the many organisations and volunteers of indigenous language communities in Nepal who gave us their assistance. The research was conducted in association with the Centre for Nepal and Asian Studies (CNAS) at Kirtipur under the Bilateral Agreement for Academic Cooperation between Tribhuvan University and Leiden University. As part of the European Science Foundation EUROCORES Programme OMLL, this work was supported by the Arts and Humanities Research Council and the EC Sixth Framework Programme under Contract no. ERAS-CT-2003980409. T.K., G.L.v.D. and P.dK. were supported by funds from the Netherlands Organisation for Scientific Research (NWO

E.J. Parkin et al. / Forensic Science International 166 (2007) 176–181

grant number 231-70-001). M.A.J. was supported by a Wellcome Trust Senior Fellowship in Basic Biomedical Science (grant no. 057559). References [1] L. Roewer, M. Krawczak, S. Willuweit, M. Nagy, C. Alves, A. Amorim, K. Anslinger, C. Augustin, A. Betz, E. Bosch, A. Caglia`, A. Carracedo, D. Corach, T. Dobosz, B.M. Dupuy, S. Fu¨redi, C. Gehrig, L. Gusma˜o, J. Henke, L. Henke, M. Hidding, C. Hohoff, B. Hoste, M.A. Jobling, H.J. Ka¨rgel, P. de Knijff, R. Lessig, E. Liebeherr, M. Lorente, B. Martı´nezJarreta, P. Nievas, M. Nowak, W. Parson, V.L. Pascali, G. Penacino, R. Ploski, B. Rolf, A. Sala, U. Schmidt, C. Schmitt, P.M. Schneider, R. Szibor, J. Teifel-Greding, M. Kayser, Online reference database of Ychromosomal short tandem repeat (STR) haplotypes, Forensic Sci. Int. 118 (2001) 103–111. [2] P. Lincoln, A. Carracedo, Publication of population data of human polymorphisms, Forensic Sci. Int. 110 (2000) 3–5. [3] T. Kraayenbrink, P. de Knijff, G.L. van Driem, J.R.M.L. Opgenort, M.A. Jobling, E.J. Parkin, C. Tyler-Smith, D.R. Carvalho-Silva, K. Tshering, G. Barbujani, I. Dupanloup, G. Bertorelle, N.M. Tuladhar, Language and Genes of the Greater Himalayan Region, 2006 OMLL Volume http:// www.le.ac.uk/genetics/maj4/Himalayan_OMLLreport.pdf. [4] G.L. van Driem, Languages of the Himalayas: An Ethnolinguistic Handbook of the Greater Himalayan Region, Containing an Introduction to the Symbiotic Theory of Language, 2 vols., Brill, Leiden, 2001. [5] E.J. Parkin, T. Kraayenbrink, G.L. van Driem, K. Tshering, P. de Knijff, M.A. Jobling, 26-locus Y-STR typing in a Bhutanese population sample, Forensic Sci. Int. (2005).

181

[6] Y Chromosome Consortium, A nomenclature system for the tree of human Y-chromosomal binary haplogroups, Genome Res. 12 (2002) 339–348. [7] J.M. Butler, R. Schoske, P.M. Vallone, M.C. Kline, A.J. Redd, M.F. Hammer, A novel multiplex for simultaneous amplification of 20 Y chromosome STR markers, Forensic Sci. Int. 129 (2002) 10–24. [8] E. Bosch, A.C. Lee, F. Calafell, E. Arroyo, P. Henneman, P. de Knijff, M.A. Jobling, High resolution Y chromosome typing: 19 STRs amplified in three multiplex reactions, Forensic Sci. Int. 125 (2002) 42–51. [9] L. Gusma˜o, J.M. Butler, A. Carracedo, P. Gill, M. Kayser, W.R. Mayr, N. Morling, M. Prinz, L. Roewer, C. Tyler-Smith, P.M. Schneider, DNA Commission of the International Society of Forensic Genetics (ISFG): an update of the recommendations on the use of Y-STRs in forensic analysis, Forensic Sci. Int. (2006) 187–197. [10] S. Schneider, D. Roessli, L. Excoffier, Arlequin ver. 2.0: A software for population genetics data analysis, 2.0 ed., Genetics and Biometry Laboratory, University of Geneva, Geneva, Switzerland, 2000. [11] H.-J. Bandelt, P. Forster, A. Ro¨hl, Median-joining networks for inferring intraspecific phylogenies, Mol. Biol. Evol. 16 (1999) 37–48. [12] R. Qamar, Q. Ayub, A. Mohyuddin, A. Helgason, K. Mazhar, A. Mansoor, T. Zerjal, C. Tyler-Smith, S.Q. Mehdi, Y-chromosomal DNA variation in Pakistan, Am. J. Hum. Genet. 70 (2002) 1107–1124. [13] J.M. Butler, Recent developments in Y-short tandem repeat and Y-single nucleotide polymorphism analysis, Forensic Sci. Rev. 15 (2003) 91–111. [14] W. Lattanzi, M.C. Di Giacomo, G.M. Lenato, G. Chimienti, G. Voglino, N. Resta, G. Pepe, G. Guanti, A large interstitial deletion encompassing the amelogenin gene on the short arm of the Y chromosome, Hum. Genet. 116 (2005) 395–401. [15] K. Thangaraj, A.G. Reddy, L. Singh, Is the amelogenin gene reliable for gender identification in forensic casework and prenatal diagnosis? Int. J. Legal. Med. 116 (2002) 121–123.

Forensic Science International 166 (2007) 182–189 www.elsevier.com/locate/forsciint

Succession pattern of carrion-feeding insects in Paramo, Colombia Efrain Martinez, Patricia Duque, Marta Wolff * Grupo interdisciplinario de Estudios Moleculares (GIEM). Universidad de Antioquia. AA, 1226 Medellı´n, Colombia Received 8 April 2004; accepted 10 May 2006 Available online 21 June 2006

Abstract The minimum postmortem interval can be estimated based on knowledge of the pattern of insect succession on a corpse. To use this approach requires that we take into account the rates of insect development associated with particular climatological conditions of the region. This study is the first to look at insect succession on decomposing carcasses in the high altitude plains (Paramo) in Colombia, at 3035 m above sea level. Five stages of decomposition were designated with indicator species identified for each stage: Callı´phora nigribasis at the fresh stage; Compsomyiops verena at the bloated stage; Compsomyiops boliviana during active decay; Stearibia nigriceps and Hydrotaea sp. during advanced decay and Leptocera sp. for dry remains. A succession table is presented for carrion-associated species of the region, which can be used for estimating time since death in similar areas. Compsomyiops boliviana is reported for the first time in Colombia. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Forensic entomology; Paramo; Insect succession; Neotropics

1. Introduction Forensic entomology is a frequently used tool to estimate the time interval between death and the discovery of the body. This period is known as the postmortem interval or PMI. For intervals greater than 72 h, forensic entomology can be more accurate in determining PMI than traditional techniques, and sometimes is the only available method [1]. The carrion-feeding invertebrate fauna is principally made up of insects and the estimation of PMI using the succession of insects on cadavers and age necrophagous larvae has been used extensively [2–8]. The first records of applied forensic entomology occurred in China in the 13th Century [9]. Forensic entomology was used for the first time as a legal instrument in a court of law in France in 1850 [10]. Since then the number of studies carried out in this field has increased significantly, particularly in Europe, United States, Canada and in some parts of the subtropics [1,2,10,11–16]. In the Neotropics, relatively few studies have been carried out [17,18,3,4,19,6,20–22]. In countries such as Colombia, environmental conditions and regional climatic conditions can

* Corresponding author at: Instituto de Biologı´a, Universidad de Antioquia, A.A. 1226, Medellı´n, Colombia. Tel.: +57 4 210 5662; fax: +57 4 233 01 20. E-mail address: [email protected] (M. Wolff). 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.027

change drastically over short distances. In any study of this type we would expect to find a similar process of insects being involved in the recycling of cadavers, but the occurrence of the species involved in this process should vary throughout different climactic regions. To date most of the neotropical studies have been done in warm tropical conditions, and none has looked at the process at high altitudes. This study addresses our lack of knowledge in this area as part of a larger ongoing effort to identify the carrion-feeding entomofauna and associated insect succession patterns at altitudes ranging from 0 to >3000 m above sea level. 2. Materials and methods The study was carried out in the Paramo region of Chingaza National Park, located in the Eastern range of the Colombian Andes, (738300 to 738550 W; 48200 to 48500 N) at an altitude of 3035 m above sea level. Average annual temperatures fluctuate between 4.5 and 21.4 8C and the annual relative humidity is above 80% [23]. The landscape of the Paramo is characterized by the presence ‘‘ fraylejones’’ (Espeletia spp.), ‘‘chuscales’’ (Chusquea tessellate) and ’’pajonales (Calamagrostis spp.) [24]. Three pigs (Sus scrofa L.) were used as models. Each pig weighed approximately 10 kg and was killed by cardiac puncture on September 12, 2002. Immediately after death, the pigs were placed in individual metal cages (60 cm  40 cm  40 cm)

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

183

Table 1 Succession of insects associated with exposed carcasses on Paramo. Ecological category

Order

Necrophagous

Diptera

Family

Calliphoridae

Sarcophagidae

Predators

Diptera Coleoptera

Tachinidae Histeridae Carabidae Staphylinidae

Species

Calliphora nigribasis Compsomyiops boliviana Sarconesiopsis magellanica Compsomyiops verena Helicobia sp. Microcerella sp Unident

Melyridae Dytisidae

Unident Neopachylopus sp. Carabus sp. Lathropinus sp. Dianous sp. Anacyptus sp. Stenus sp. Dasyrhadus sp. Copelatus sp.

Sphecidae

Fresh

Bloated

Active

L

E

A

L

E

A

L

E

A

L

P

A

L

P

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x x

x

x

x x

x

x

x

x x x x x x

x x x x x

x

x

Ichneumonidae Braconidae Encyrtidae Chalcididae

Unident Unident Unident Unident

x

Forficulidae Enicocephalidae Reduviidae Piophilidae

Coleoptera

Sepsidae Drosophilidae Mycetophilidae Ptillidae

Unident Unident Unident Stearibia nigriceps Azelia sp. (probably) Dasymorellia seguyi Fannia sp. Helina sp. Hydrotaea sp. Limnophora sp. Syllimnophora atroviatta Muscina stabulans Leptocera sp. Beckerina sp. Dohrniphora sp. Metopina sp. Diploneura sp. Borophaga sp. Unident Orygma sp. Drosophila sp. Speolepta sp. Actinopteryx sp.

Omnivorous

Hymenoptera Dyctioptera

Vespidae Blattellidae

Unident Unident

Incidental

Diptera

Anthomyzidae Camillidae Ceratopogonidae Chamaemyiidae

Ischnomyia sp. Unident Unident Unident

Dermaptera Hemiptera

Saprophagous

Diptera

Muscidae

Sphaeroceridae Phoridae

x

x

x

Unident

Predators

x

x x

x x

Hymenoptera

Remains

A

x

Predators/ parasitois

Advanced

x

x x x x x

x x x x x

x x

x

x

x

x x

x

x x x x x

x

x x x

x

x

x

x

x

x

x

x x

x

x

x

x x

x x x

x x x

x x x

x

x

x

x x x x

x x x x

x

x

x

x x

x

x x

x x

x x x

x x

x x x x

x

x x

x

x x x x x

x

x

x

x x

x x

x x x

x

x

x

x

184

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

Table 1 (Continued ) Ecological category

Order

Family

Species

Fresh A

Coleoptera

Hemiptera

Chloropidae Dolycopodidae Otitidae Psycodidae Sciaridae Sciomycidae Syrphidae Tabanidae Cantaridae Chrysomelidae Coccinelidae Curculionidae Lampyridae Lycidae Melolontidae Nitidulidae Oedemeridae Cercopidae Cicadelidae Lygaeidae Miridae Pentatomidae Pyrrhocoridae

Meromyza sp. Unident Cephalia sp. Psychoda sp. Unident Unident Eristalis sp. Tabanus sp. Cantharis sp. Unident Unident Unident Pollaclasis sp. Photuris sp. Unident Unident Conotelus sp. Unident Unident Unident Unident Unident Unident Unident

Bloated L

E

A

L

Active E

A

x x x

x x x x x x

x x x

x x

x

x

L

Advanced E

A

L

Remains P

A

x x

x x

x

x

x x x

L

P

x x x

x x

x

x x x

x

x

x x

x x

x x x x

x x x x x

x

A: Adult; L: Larvae; E: Eggs; P: Pupe.

made with 2 cm  2 cm wire meshing. This allowed access to the carcasses by insects while preventing disturbance by vertebrate scavengers [6]. The three cages were placed 50 m from each other. The cadavers were observed for 8 h immediately following sacrifice. Subsequently, the process of decomposition was observed during daylight hours from September 12 to December 4 2002. Samples were taken three times a day (7.00, 13.00 and 19.00) for the first eight days; twice a day (at 11.00 and 17.00) for the next eight days and subsequently once a day (at 12.00) until the dry remains stage had been reached 83 days after death. At each sampling time, arthropods flying over or perched on the carcass were collected with an entomological net before the carcass was moved. Then, insects in natural cavities (eyes, nose, mouth, anus) and the cardiac puncture wound were collected. Finally, three times a day the insects were collected underneath the carcass and in the soil to a depth of 10 cm. Immature specimens were fixed in 80% alcohol and the adults were killed with ethyl acetate and mounted with entomological pins. Ambient and body (rectal) temperature were taken at the same time as collecting the specimens. Observations were made on the physical changes of the carcass over time (colour, degree of swelling, discharge of liquids and gases, etc). The carcass was weighed once a day with a scale [25]. The first eggs mass founded, the colonising species, were identified by collecting eggs from the carcasses. These were placed in plastic containers (three per pig) containing approximately 200 g of raw pig liver. The containers were immediately covered with muslin and placed in styrofoam boxes. The development of the insects that hatched was monitored until the adult stage was reached [26].

The taxonomic identification of the adults and larvae was carried out using the following keys; Greenberg and Szyska [18,21], Borror et al. [27], Carvalho [28], Dear [29], Liu and Greenberg [30], Mariluis [31], Mc Alpine et al. [32,33], Queiroz and Carvalho [34], Smith [2,35], Stehr [36], Wells et al. [37] and White [38]. When necessary, the larvae were cleared with KOH and permanently mounted in Canada balsam. All the specimens were deposited in the Universidad de Antioquia Entomological Collection. 3. Statistical analysis This study was designed to evaluate the succession pattern of insects during the decomposition of the three cadavers, and was not designed for a quantitative analysis. However, insect species were assigned a number and the presence or absence of each species on each cadaver on each day was recorded and the data analysed with a Kruskal–Wallis test P > 0.05. 4. Results Flies (dipterans) and beetles (coleopterans) constituted 99.6% of the individuals collected. There were no significant differences between the three cadavers used with respect to the presence and absence of species on each day (Kruskal–Wallis test, P > 0.05). The data then were combined to define the succession pattern. In total, 36,892 individuals were collected (larvae, nymphs and adults) belonging to 6 orders, 53 families, 42 genera and 98 species and morph species (Table 1). Of these, 34,832 (94.4%) were flies (Diptera), 1,881 (5.1%) beetles (Coleoptera), 73

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

185

Fig. 1. Daily temperature variations related with decomposition phase.

Fig. 2. Daily weight variations related with decomposition phase.

(0.2%) bugs (Heteroptera), 73 (0.2%) wasps (Hymenoptera), 26 (0.07%) earwigs (Dermaptera) and 7 (0.02%) cockroaches (Dyctiopetera). Within the Diptera two families were predominant: blowflies (Calliphoridae) (91% of all Diptera) and small dung flies (Sphaeroceridae) (6.4% of the Diptera). The other 2.6% corresponded to individuals belonging to another 20 families. Of the coleopterans, four of the 15 families represented were most abundant: featherwing beetles (Ptillidae) (46%), rove beetles (Staphylinidae) (24%), hister beetles (Histeridae) (19%) and soft winged flower beetles (Melyridae) (8.6%). According to the classification of ecological categories by Smith [2] and Magan˜a [39], the following groups were present among the entomofauna collected: necrophagous species, predators, parasites of necrophagous species, saprophagous species, omnivorous species and incidental or local fauna (Table 1).

remains (Table 1), which were determined by the physical and body temperature changes of the carcass [6].

5. Stages of decomposition Five different stages of decomposition were observed during the study: fresh, bloated, active decay, advanced decay and dry

5.1. Fresh (0–3 days) A sharp decrease in body temperature was observed reaching levels below ambient temperatures (Fig. 1). Oviposition was observed in natural orifices, on the neck, around the wound and on the side of the carcass touching the ground. Immediately after death, the first adults of Muscidae (Dasymorellia seguyi, Fannia sp., Limnophora sp.) and Calliphoridae (Calliphora nigribasis and Compsomyiops verena) arrived. From 4 h after death and until day 18, Calliphora nigribasis were observed ovipositing. Three days after death, 1st instar larvae of Calliphora nigribasis were collected in natural orifices and around the wound (Table 3 and Appendix A). The second species to oviposit was Compsomyiops verena from day 3 until day 7. Other abundant adult insects during this stage were Leptocera sp., Ichnomyia sp. and coleopterans Actinopteryx sp. and Dianous sp.

Table 2 Percentage of adult dipterans collected at different stages of decomposition Family

Species

Fresh

Bloated

Active

Advanced

Remains

Calliphoridae

Calliphora nigribasis Compsomyiops verena Sarconesiopsis magellanica

31.08 1.35 0.00

33.80 20.21 1.38

26.03 42.47 2.05

0.21 0.00 0.00

0.27 0.14 0.00

Muscidae

Limnophora sp. Dasymorellia seguı´ Muscina stabulans Syllimnophora atroviatta Fannia sp. Hydrotaea sp.

16.22 6.76 4.05 4.05 2.70 0.00

3.05 1.66 17.17 3.32 0.28 0.83

1.37 4.11 12.33 0.68 0.00 3.42

0.07 0.14 0.21 0.34 0.00 0.34

0.14 0.00 0.00 0.00 0.00 0.00

Sarcophagidae

Helicobia sp. Microcerella sp.

1.35 0.00

0.55 0.28

0.00 0.00

0.00 0.00

0.00 0.00

Piophilidae Sphaeroceridae Anthomyzidae Syrphidae Tachinidae Dolychopodidae Sciaridae

Stearibia nigriceps Leptocera sp. Ischnomyia sp. Eristalis sp. Unident Unident Unident

0.00 20.27 12.16 0.00 0.00 0.00 0.00

0.00 13.30 1.11 0.83 1.67 0.00 0.28

0.00 3.42 0.68 0.68 2.68 0.00 0.00

0.69 95.54 0.00 0.07 0.14 1.78 0.48

0.41 98.10 0.00 0.00 0.00 0.81 0.14

186

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

5.2. Bloated (4–16 days)

Table 3 Percentage of dipteran larvae activity at each stage of decomposition

Body temperature began increasing during this stage (Fig. 1) as a result of insect activity and the putrefaction process [40]. The carcass lost 10% of its initial weight (Fig. 2). Abundant masses of larvae were observed throughout the carcass. First and 2nd instar larvae belonging to Calliphoridae, Sarcophagidae and Muscidae were found as well as 3rd instar larvae of Calliphora nigribasis, Compsomyiops verena, Muscidae (Azelia sp. probably) and Sarcophagidae (Appendix A). A large number of adults of Calliphora nigribasis, Compsomyiops verena, Leptocera sp. and Muscina stabulans were attracted to the carcass during this stage (Table 2). Ovipositing individuals of these species were also collected. A surge in predators was evident at this stage. New groups of coleopterans appeared: Neopachylopus sp., Dasyrhadus sp.; Anacytyus sp., Dianous sp., Lathropinus sp., Carabus, in addition to other groups of dipterans, dermapterans, hymenopterans and hemipterans (Table 1).

Family

Species

Bloated

Active

Advanced

Remains

Calliphoridae

Callı´phora nigribasis Compsomyiops boliviana Compsomyiops verena

89.24

81.55

51.07

31.89

0.00

8.23

44.41

62.52

10.14

8.77

1.43

0.17

Helina sp. Hydrotaea sp. Azelia sp. (probably) Muscina sp.

0.00 0.00 0.51

0.00 0.00 1.27

0.06 0.48 2.20

0.00 1.37 0.70

0.00

0.18

0.01

0.00

Stearibia nigriceps Leptocera sp. Indent Indent

0.00

0.00

0.32

2.01

0.00 0.10 0.00

0.00 0.00 0.00

0.00 0.00 0.02

1.34 0.00 0.00

5.3. Active decay (17–30 days) The temperature of the cadaver remained low and only increased in the last days of active decay (Fig. 1). A sharp and rapid loss of weight occurred (Fig. 2) in which 52% of the total body weight was lost. With respect to the insects present, all three larval instars of Calliphora nigribasis and 2nd and 3rd instar larvae of Compsomyiops boliviana and 3rd instar larvae of Compsomyiops verena were found. No individuals of the family Sarcophagidae were observed (Table 1). Larval activity mainly took place in large masses concentrated towards the side of the carcass touching the ground. This activity continued to be dominated by Calliphora nigribasis and Compsomyiops verena. Ovipositing specimens of the latter species were collected on days 18 and 19 (Table 2). Other predators present were individuals of hymenoptera (Ichneumonidae and Vespidae) and coleoptera (Staphylinidae) (Table 1). 5.4. Advanced decay (31–51 days) Peaks in cadaver temperature were reached during this stage with a maximum value of 100.40 8F (Fig. 1). By the end of advanced decay, the body had lost 75% of its initial weight (Fig. 2) and a mucilaginous material was found on and around the cadaver. This by product of decomposition (BOD) consists some of the internal tissues with insect material and other products of decomposition [41]. Larval activity in the carcass was notably reduced although it continued to be dominated by Calliphoridae (Table 3). The majority of the larval masses were concentrated in the muddy ground under the body. Several dead larvae were found on the body, underneath it and in the immediate vicinity. We colleted Calliphora nigribasis and Compsomyiops boliviana larvae predating individuals of its own species (intra-specific

Muscidae

Piophilidae Sphaeroceridae Sarcophagidea Tabanidae

predation) and of the other species (inter-specific predation). New saprophagous species were first observed at this stage such as: Stearibia nigriceps (Piophilidae), collected after day 40 and Helina sp. and Hydrotaea sp. (Muscidae), collected after day 34 (Appendix A). Pupae of Calliphora nigribasis, Compsomyiops boliviana and some Muscidae were found in the remains of the carcass. Pupae of Calliphora nigribasis and Compsomyiops verena were collected at a distance of 2 m from the carcass among layers of moss on the ground. These were placed in rearing containers and raised to adults (Table 1). The number of adult Calliphoridae dropped considerably, being replaced by small dipterans such as Leptocera sp. and Stearibia nigriceps which may have been attracted by the decomposing tissues and the BOD (Table 2). 5.5. Dry remains (52–83 days) The last collections were made during this stage. By the end, 83% of the initial body weight had been lost (Fig. 2). Only remains of skin and bones were left. The BOD had dried and some of this had mixed with the soil. The small numbers of larvae found were concentrated in the extremities (hooves), places providing shelter and where small amounts of soft tissue remained. Second instar larvae of Calliphora nigribasis and Compsomyiops boliviana were found until days 57 and 64, respectively. Third instar larvae and pupae were found until the end, as was the case with 3rd instar larvae of Stearibia nigriceps (Appendix A). 6. Discussion Five stages of decomposition of the cadaver were observed: fresh, bloated, active decay, advanced decay and dry remains. The different stages were categorized according to changes in appearance and body temperature. The fresh stage lasted four days and the bloated stage lasted 13 days. This was longer that the corresponding periods reported in other studies [1,6,41],

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

where the fresh stage lasted for 2 days and the bloated stage lasted from 4 to 8 days. In our study, subsequent stages conformed to the time periods given in the aforementioned studies. This situation may be explained by the particular conditions of cooler temperatures in our studies compared with the other reports. Calliphora nigribasis and Compsomyiops verena were the first species to colonise the carcasses. This differed from the study carried out by Wolff et al. [6] in Colombia at 1450 m above sea level where Phaenicia sericata was the first species to arrive. During the bloated stage and active decay, Calliphora nigribasis, Compsomyiops boliviana and Compsomyiops verena were mainly responsible for the loss of carcass tissue. The presence of a large number of individuals of these species is the main reason for the rapid loss of weight during advanced decay and consequently a reduced availability of food (Table 2, Fig. 2). This competition for resources may, in turn, may have contributed to the interspecific predation between larvae of Calliphora nigribasis and Compsomyiops boliviana. The succession pattern observed shows similarities between those described by Anderson and VanLaerhoven [1], Wolff et al. [6] and Tullis and Goff [41], in those two principal groups (Diptera and Coleoptera) were represented. Diptera, especially Calliphora nigribasis and Compsomyiops boliviana continued their necrophagous activity during the whole decomposition process (Table 3). which normally is not the case in most reports in which Calliphoridae are present only up to the first 3–4 stages of decomposition, but are not found on dried. The environmental conditions of the study area, at an altitude of 3035 m and with environmental temperatures fluctuating between 9 and 23 8C during the day and 1 to 7 8C at night, influenced the process of decomposition. The decomposition process of the carcasses took 83 days as opposed to the 45 days reported by Carvalho and Linhares [19] in Brazil during the wet season where the same 10 kg model was used. Although the weight of the pigs was equal, the rate of decomposition may have been influenced by the environmental temperature. This study was done at 3035 m above sea level, whereas the study by Castillo-Miralbes [42] was done at 300 m above sea level and at higher ambient temperatures. The rapid weight loss of the cadavers was a result of the conversion of carcass biomass into larval biomass and the subsequent exit of insects from the body during pupation [41]. The percentage of carrion removed over time was much less than a study carried out in Hawaii [41] with pigs of a similar size. At the end of the active stage in Hawaii, 80% of body weight had been lost and subsequently only a small reduction was observed. At the end of this study [41] 18% of the body weight remained, whereas in our study only 52% of the body weight had been removed by the end of the active stage. At the advanced stage, 75% had been lost and only 13% was left at the dry remains stage (Fig. 2). This may be explained by the marked difference in the ovipositing habits of the dipterans. Calliphora nigribasis and Compsomyiops verena were

187

observed ovipositing until the first few days of the active stage while in Hawaii, Chrysomya megacephala and Chrysomya rufifacies were only observed ovipositing during the fresh stage. This late oviposition could also explain the presence of 2nd instar larvae of Calliphora nigribasis and Compsomyiops boliviana during the dry remains stage and 3rd instar larva of the same species being present until the last day of sampling (Appendix A). Larvae of Piophilidae (Stearibia nigriceps = P. foveolata) and Muscidae (Hydrotaea sp.) were founded in advanced and remains. Stearibia nigriceps important forensic indicator because arrives frequently after saponification in the last decomposition stages [1,43]. In the present and other studies, at Colombia, Stearibia nigriceps presented the same behaviour [6,22]. The individuals identified as Azelia sp. and Helina sp., following Smith [35], were all larvae. Due to the early age stages and the lack of a larvae key for this group, we do not give a definitive identification for these individuals. In this region, and with the specific environmental conditions, we can designate indicator species for the different stages of decomposition, which were determined by the physical and body temperature changes of the carcass and based on the presence of the species for the first time on each of the different decomposition stages, as follows: fresh Calliphora nigribasis; bloated – Compsomyiops verena; active decay – Compsomyiops boliviana; advance decay – Stearibia nigriceps and Hydrotea sp.; dry remains – Leptocera sp. Although indicator species are not used for the estimation of PMI, they are useful to know the stage of decay when combinations of species occur. This is the first report of Compsomyiops boliviana in Colombia. 7. Conclusions Altitude, bio-climatic factors and vegetation determine species distribution in the tropics. This is clearly shown by this study carried out on Paramo at 3035 m above sea level where the Andean species, Calliphora nigribasis, Compsomyiops boliviana and Compsomyiops verena were found. These species had not previously been reported in studies of forensic entomology in Colombia, conducted in different climactic zones. These species may be important indicator species of specific stages of decomposition in this bio-climatic zone. Finally, a succession table is presented for insects on a carcass in Paramo. This aims to be a useful instrument for other researchers and forensic scientists working in similar areas (Appendix A). Acknowledgements This study received financial support from Colciencias (Instituto Colombiano para el Desarrollo de la Ciencia y la Tecnologı´a) project number 1115-05-11503 and from the Universidad de Antioquia, Medellı´n, Colombia and from Drs Jeffrey Wells and Carl Lowenberger for the revision of manuscript.

188

E. Martinez et al. / Forensic Science International 166 (2007) 182–189

Appendix A Succession table (presence–absence) Species

A

B C D E F G H I

J

Dev

Fresh

L1 L2 L3 L2 L3 L3 L2 L3 L3 L2 L3 L3 L2 L3 L1 L2 L3 L3

Bloated

Active

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

27

29

30

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0

1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0

1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0

1 1 0 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0

1 1 1 0 0 1 1 0 0 0 0 0 0 0 0 1 1 0

1 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0

1 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0

1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0

1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 0 0 0 1 0 0 0 0 0 0

1 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 0 0 0 0 0 1 0 0 0 0 0 0

1 1 1 1 1 1 0 1 0 0 0 1 0 0 0 0 0 0

1 1 1 1 1 1 0 1 0 0 0 1 0 0 0 0 0 0

0 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0

Species Dev Advanced

Remains

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 A

B C D E F G H I

J

L1 L2 L3 L2 L3 L3 L2 L3 L3 L2 L3 L3 L2 L3 L1 L2 L3 L3

1 1 1 1 1 1 1 1 0 0 0 1 0 0 0 0 0 0

0 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0

0 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0

0 1 1 1 1 1 0 1 0 1 0 0 0 0 0 0 0 0

0 1 1 1 1 1 0 1 1 0 1 0 0 0 0 0 0 0

0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0

0 1 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0

0 1 1 1 1 1 0 1 0 0 1 0 0 0 0 0 0 0

0 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0

0 1 1 1 1 0 1 1 0 0 1 0 0 1 0 0 0 0

0 1 1 1 1 0 1 1 0 0 1 0 0 0 0 0 0 0

0 1 1 1 1 0 1 1 0 0 1 0 0 0 0 0 0 0

0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 0 0

0 1 1 1 1 0 1 1 0 1 1 0 0 0 0 0 0 0

0 1 1 1 1 1 1 1 0 0 0 0 0 1 0 0 0 0

0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0

0 1 1 1 1 1 0 1 0 1 1 0 1 1 0 0 0 0

0 1 1 1 1 1 0 1 0 1 0 0 0 1 0 0 0 0

0 0 1 1 1 1 0 1 1 0 1 0 1 1 0 0 0 0

0 1 1 1 1 0 1 1 0 1 1 0 1 1 0 0 0 0

0 1 1 1 1 1 0 1 0 1 1 0 0 1 0 0 0 0

0 1 1 1 1 1 0 1 0 0 1 0 0 1 0 0 0 1

0 1 1 1 1 1 0 1 0 0 0 0 0 1 0 0 0 0

0 0 1 1 1 1 0 1 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 1 0 1 0 1 1 0 0 1 0 0 0 0

0 1 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 0 0 0 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 1

0 0 1 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0

0 0 1 1 1 0 0 1 0 0 1 0 0 0 0 0 0 0

0 0 1 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0

0 0 1 1 1 0 0 0 0 0 1 0 0 1 0 0 0 0

0 0 1 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0

0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0

0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 0 0 1

L: larva, Species A: Calliphora nigribasis, B: Compsomyiops boliviana, C: Compsomyiops verena, D: Azelia sp., E: Helina sp., F: Hydrotaea sp., G: Muscina sp., H: Stearibia nigriceps, I: Sarcophagidae (unidentified), J: Leptocera sp.

References [1] G. Anderson, S.L. VanLaerhoven, Initial studies on insect succession on carrion in Southwestern British Columbia, J. Forensic Sci. 41 (1996) 617–625. [2] K.G.V. Smith, A Manual of Forensic Entomology, Trustees of the British Museum (Natural History) and Cornell University Press, London, 1986, p. 205. [3] R. Ferllini, Determinacio´n del tiempo de muerte en cada´veres putrefactos, momificados y saponificados, Medicina Legal de Costa Rica 10 (2) (1994) 17–20. [4] M.O. Moura, C.J.B. Carvalho, E. Monteiro, A preliminary analysis of insects of medico-legal importance in Curitiba, State of Parana, Mem. Inst. Oswaldo Cruz 92 (2) (1997) 269–274.

[5] M. Wolff, Primeros Estudios de Entomologı´a Forense en Medellı´n: Presentacio´n de Algunos Casos, in: C.A., Giraldo (Ed.), Casos Forenses en Medicina Legal, 13 (2001) 25-34. [6] M. Wolff, A. Uribe, A. Ortiz, P. Duque, A Preliminary study of forensic entomology in Medellı´n, Colombia Forensic Sci. Int. 120 (1/2) (2001) 53–59. [7] B. Gremberg, J. Wells, Forensic use of Megaselia abdita and M. scalaris (Phoridae: Diptera): case studies development rates and eggs structure, J. Med. Entomol. 35 (3) (1998) 205–209. [8] K.L. Sukontanson, K. Sukontanson, P. Narongchai, S. Lertthamnongtham, S. Piangjai, J:K. Olson, Chrysomya rufifacies (Macquart) as a forensically fly species in Thailand: a case report, J. Vector Ecol. 26 (2) (2001) 162–164. [9] M. Benecke, A brief history of forensic entomology, Forensic Sci. Int. 120 (2001) 2–14.

E. Martinez et al. / Forensic Science International 166 (2007) 182–189 [10] M.L. Goff, Estimation of postmortem interval using arthropod development and successional pattern, Forensic Sci. Int. 5 (1993) 82–94. [11] J. Wells, B. Greenberg, Resource use by an introduced and native carrion flies, Oecologie 99 (1994) 181–187. [12] M.L. Goff, M.M. Flynn, Determination of postmortem interval by arthropod succession: A case study from the Hawaiian Islands, J. Forensic Sci. 36 (1991) 607–614. [13] M.L. Goff, Comparison of insects species associated with decomposed remains recovered inside dwellings and outdoors on the island Oahu Hawaii, J. Forensic Sci. 36 (1991) 748–753. [14] M.L. Goff, A.I. Omori, K. Gunatilake, Estimation of postmortem interval by arthropod succession. Three case studies from Hawaiian Islands, Am. J. Forensic Med. Pathol. 9 (1988) 220–225. [15] M.L. Goff, C.B. Odom, Forensic entomology in the Hawaiian Islands. Three case studies, Am. J. Forensic Med. Pathol. 8 (1) (1987) 45–50. [16] A. Oliva, Insects of forensic significance in Argentina, Forensic Sci. Int. 120 (1/2) (2001) 145–154. [17] L.F. Jiron, V.M. Cartin, Insect succession in the decomposition of a mammal in Costa Rica, New York Entomol. Soc. 89 (1981) 158–165. [18] B. Greenberg, M.L. Szyska, Immature stages biology of fifteen species of Peruvian Calliphoridae (Diptera), Ann. Entomol. Soc. Am. 77 (1984) 488–517. [19] L.M.L. Carvalho, A.X. Linhares, Seasonality of insect succession and pig carcass decomposition in a natural forest area in Southeastern Brazil, J. Forensic Sci. 46 (3) (2001) 604–608. [20] M. Barreto, M.E. Burbano, P. Barreto, Files (Calliphoridae, Muscidae) and beetles (Silphidae) from human cadavers in Cali, Colombia Mem. Inst. Oswaldo Cruz. 97 (1) (2002) 137–138. [21] B. Greenberg, M.L. Szyska, Key to know third instar larvae of Peru (modified from B. Greenberg and M.L. Szyska 1984), in: B. Greenberg, J.C. Kunich (Eds.), Entomology and the Law Flies of Forensic Indicators, Cambridge University, London, 2002, p. 94. [22] S. Perez, P. Duque, M. Wolff, Successional behavior occurrence matrix of carrion-associated arthropods in the Urban Area of medellin, Colombia J. Forensic Sci. 50 (2) (2005) 448–454. [23] Instituto Nacional de Recursos Naturales Renobables y del Ambiente INDERENA. Parque Nacional Natural Chingaza, Santafe´ de Bogota´, Colombia, 1986. p. 255. [24] P. Pedraza-Pen˜alosa, J. Betancur, P. Franco-Rosselli, Chisaca´, Un Recorrido Por Los Pa´ramos Andinos, Instituto de Ciencias Naturales, U. Nacional de Colombia. Instituto de Investigacio´n de Recursos Biolo´gicos Alexander von Humboldt. Colombia, 2004. p. 13. [25] N.H. Haskell, W.D. Lord, J.H. Byrd, Collection of entomological evidence during death, in: J.H. Byrd, J.L. Castner (Eds.), Forensic Entomology, CRC Press, United States of America, 2001, pp. 81–120.

189

[26] J.H. Byrd, Laboratory rearing of forensic insects, in: J.H. Byrd, J.L. Castner (Eds.), Forensic Entomology, CRC Press, United States of America, 2001, pp. 121–142. [27] D. Borror, C. Triplehorn, N. Johnson, An Introduction to the Study of Insects, 6th ed., Sauders College Publishing, United States of America, 1989, p. 875. [28] C. J. B. Carvalho, Muscidae (Diptera) of the Neotropical region: Taxonomy. Editora UFPR, Brasil, 2002. p. 287. [29] J.P. Dear, A revision of the new world Chrysomyini (Diptera: Calliphoridae), Rev. Bras. Zool. 3 (3) (1985) 109–169. [30] D. Liu, B. Greenberg, Immature stages of some flies of forensic importance, Ann. Entomol. Soc. Am. 82 (1) (1989) 80–93. [31] J.C. Mariluis, Key to common adult blowflies of South America, in: B. Greenberg, J.C. Kunich (Eds.), Entomology and the Law Flies of Forensic Indicators, Cambridge University, London, 2002, pp. 94–120. [32] J.F. Mc Alpine, B.V. Peterson, G.E. Shewell, H.J. Teskey, J.R. Vockeroth, D.M. Wood, Manual of Nearctic Diptera, vol. 1, Minister of Supply and Services Quebec, Canada, 1981, p. 1332. [33] J.F. Mc Alpine, B.V. Peterson, G.E. Shewell, H.J. Teskey, J.R. Vockeroth, D.M. Wood, Manual of Nearctic Diptera, vol. 2, Minister of Supply and Services Quebec, Canada, 1987, p.1332. [34] S.M.P. Queiroz, C.J.B. Carvalho, Chave Picto´rica e Descric¸oes de Larvas de 38 I´nstar de Diptera (Calliphoridae, Muscidae e Fanniidae) em Vazadouros de Resı´duos So´lidos Dome´sticos em Curitiba, Parana´. Ann. Soc. Entomol. 16(2) (1987) 265–288. [35] K.G.V. Smith, Introduction to the Immature Stages of British Flies, Department of Entomology British Museum (Natural History), London, 1989, p. 259. [36] F.W. Stehr, Immature Insects, vol. 2, Kendall/Hunt Publishing Company, United States of America, 1991, p. 974. [37] J. Wells, J.H. Byrd, T. Tantawi, Key to third-instar Chrysomyinae (Diptera: Calliphoridae) from carrion in the Continental United States, J. Med. Entomol. 36 (5) (1999) 638–641. [38] R.E. White, A field guide to the beetles of North America, Peterson field Guides series, New York, 1999, p. 368. [39] C. Magan˜a, La entomologı´a forense y su aplicacio´n a la medicina legal, Data de la muerte. Bol. S. E. A. 28 (2001) 49–57. [40] J.A. Payne, A summer carrion study of the baby pig Sus scrofa Linnaeus, Ecology 46 (5) (1965) 592–602. [41] K. Tullis, M.L. Goof, Arthropod succession in exposed carrion in a tropical rainforest on O’ahu Island, Hawaii, J. Med. Entomol. 24 (1987) 332–339. [42] M. Castillo-Miralbe´s, Artro´podos presentes en carron˜a de cerdos en la comarca de la Litera (Huesca), Bol. S.E.A. 28 (2001) 133–140. [43] G.F. Bornemissza, An analysis of arthropods succession in carrion and the effect of its decomposition on the soil fauna, Aust. J. Zool. 5 (1957) 1–12.

Forensic Science International 166 (2007) 190–193 www.elsevier.com/locate/forsciint

Liver histopathology of fatal phosphine poisoning Sepideh Saleki a, Farid Azmoudeh Ardalan b,*, Abdullah Javidan-Nejad c,ä b

a Legal Medicine Organization of Iran, Pathology Department, Kahrizak, Tehran, Iran Tehran University of Medical Sciences, Central Pathology Department, Imam Khomeini Hospital, Dr. Gharib St., Tehran, Iran c Legal Medicine Organization of Iran, Toxicology Department, Kahrizak, Tehran, Iran

Received 8 April 2006; received in revised form 12 May 2006; accepted 14 May 2006 Available online 27 June 2006

Abstract Two commonly used pesticides in agriculture are phosphides of aluminium and zinc. Both of these metal phosphides act through elaboration of toxic phosphine gas. The poisoning in Iran is mostly oral and suicidal. Phosphine is rapidly absorbed throughout the gastrointestinal tract after ingestion and it is partly carried to the liver by the portal vein. In this study the liver histopathology of fatal poisoning is scrutinized. A descriptive, retrospective study was performed on 38 fatal phosphine poisonings. The slides of liver specimens of the cases were retrieved and studied separately by two pathologists. The poisoning was suicidal in 33 (86.5%) of cases. Portal inflammation was negligible in 37 cases and only in one of the cases, a moderate degree of chronic inflammation accompanied by granuloma formation was observed. Major histopathologic findings were as follows: mild sinusoidal congestion; 12 cases (31.6%), severe sinusoidal congestion; 25 cases (45.8%), central vein congestion; 23 cases (60.5%), centrilobular necrosis; 3 cases (7.9%), hepatocytes nuclear fragmentation; 6 cases (15.8%), sinusoidal clusters of polymorphonuclear leukocytes; 12 cases (31.6%), and mild macrovesicular steatosis; 5 cases (13.2%). Fine isomorphic cytoplasmic vacuoles were observed in 36 cases (94.7%). These vacuoles were distributed uniformly in all hepatic zones in the majority (75%) of cases. This study reveals that the main histopathologic findings of fatal phosphine poisoning in the liver are fine cytoplasmic vacuolization of hepatocytes and sinusoidal congestion. # 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Poisoning; Phosphines; Histopathology; Liver; Fatal

1. Introduction Fumigants are pesticides with relatively high vapor pressures that allow them to effectively penetrate porous materials [1]. Aluminum and zinc phosphides are the most commonly used pesticides. Aluminum phosphide is an insecticide and zinc phosphide is a rodenticide [2], both of them are converted into hydrogen phosphine (PH3) upon hydrolysis. PH3 is the active agent which results in the inhibition of cytochrome C oxidase and subsequent generation of reactive oxygen species [3–5]. The major targets of PH3 poisoning in the human body are the lungs, heart, brain, gastrointestinal tract, kidney, and liver [3,4,6].

The major mode of poisoning in western countries is occupational inhalation of the phosphine gas [1,7]. However, in Iran (and probably India) most of the documented cases of phosphine poisoning are suicidal and due to ingestion of aluminum phosphide [6,8]. After ingestion, phosphine is rapidly absorbed throughout the gastrointestinal tract and it is partly carried to the liver by the portal vein. It is known that phosphine can cause liver dysfunction, especially after the first day of poisoning. However, there are limited studies about the histopathologic changes of liver in autopsy cases of phosphine poisoning [9]. In this study, the liver histopathology of fatal poisoning is scrutinized. Future studies may show the clinical significance of the observed morphologic findings. 2. Materials and methods

* Corresponding author. Tel.: +98 21 22583942. E-mail address: [email protected] (F.A. Ardalan). ä Deceased. 0379-0738/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2006.05.033

To evaluate the histopathologic findings in the liver in lethal phosphine poisoning, a retrospective descriptive study was

S. Saleki et al. / Forensic Science International 166 (2007) 190–193

conducted in the histopathology department of the Legal Medicine Organization of Iran; H&E stained sections of the livers of all fatal phosphine poisoning victims were retrieved from our files between March 2003 and March 2004. The slides were reviewed by two pathologists, separately. In case of any disagreements about the type or severity of histopathologic findings, the two pathologists reviewed the slides together in order to reach to an agreement. The final diagnosis of fatal phosphine poisoning was based on history of ingestion of the compound, identification of the container, garlic or decaying fish odour imparted to breath, clinical features of the poisoning, and confirmed by detection of phosphine by a qualitative silver nitrate paper test [10]. Screening for drugs (including opiates) was performed by using color tests and thin layer chromatography (TLC) plates after acidic/neutral and basic extractions of gastric contents and viscera. The analysis of trace metals (e.g., mercury, bistmuth, arsenic, antimony, and tellurium) was done by Reinsch test. A Prussian blue color test was also used for cyanide poisoning. 3. Results Thirty-eight cases of confirmed fatal phosphine poisoning were enrolled in our study, 20 of the cases were males and 18 of them were females. The mean age (and standard deviation) of cases was 28.3 (16.8) years. The poisoning was suicidal in 86.5% of our cases (33 cases). Autopsies were performed in less than 24 h after death in all of the victims and no evidence of putrefication was observed at the time of autopsy. Toxicological studies for drugs and other toxins were negative in all cases except for one case that showed a strong positive result for morphine. The quality of liver sections was appropriate in 36 cases and some degrees of autolysis were observed in only 2 cases. Nevertheless, histopathologic evaluation was feasible in

191

all cases. Portal inflammation was negligible in 37 cases and only in one of the cases, a moderate to severe degree of chronic inflammation accompanied by granuloma formation was observed. The histopathologic findings are summarized in Table 1. 4. Discussion Metal phosphides are common pesticides, which are used frequently in agriculture. Upon contact with moisture in the environment, metal phosphides (e.g., aluminum phosphide) undergo a chemical reaction yielding phosphine gas, which is the active pesticidal component [3]. Some studies have revealed that phosphine poisoning can lead to liver damage [3,6,9,11]. Abnormal liver function tests in non-fatal cases of poisoning also suggest the presence of a kind of pathology in the liver [6–9,12]. Nevertheless, the morphologic findings of liver in cases of phosphine poisoning have been studied in very limited studies with small sample size [11,13,14]. On the other hand, the forensic annals in Iran is replete with cases of phosphine poisoning due to suicidal ingestion of aluminum phosphide [6]. Therefore, in this study, we have tried to show the hepatic histopathologic changes in fatal cases of phosphine poisoning in a relatively large number of cases. In our study, the most frequent histopathologic findings have been sinusoidal congestion and fine cytoplasmic vacuolization of hepatocytes. The sinusoidal congestion is occasionally so intense that it may cause clumping of red blood cells (Fig. 1). The fine cytoplasmic vacuoles which were the most consistent finding in our cases are rather uniform in size and shape (Fig. 2). The vacuoles are distributed uniformly in all acinar zones in the majority (71.1%) of cases. Other less frequent histopathologic findings include portal edema, portal vein congestion, central vein congestion, centrilobular necrosis, nuclear fragmention (Fig. 3), clusters of polymorphonuclear leukocytes in sinusoids

Table 1 Summary of histopathologic findings Mild frequency (%) Portal inflammation Portal edema Congestion of portal tract vein Sinusoidal congestion Congestion of central vein Hepatocyte nuclear fragmentation Polymorphonuclear leukocyte clusters in sinusoids Lobular necrosis Cholestasis Vacuolization of hepatocytesd,e Macrovesicular steatosis Subcapsular hemorrhage a

0 21 26 12 23 6 10 3 0 27 5 2

(0) (55.3) (68.4) (31.6) (60.5) (15.8)b (26.3) (7.9)c (0) (71.1) (13.2)f (5.3)

Moderate to severe frequency (%) 1 0 2 25 0 0 0 0 0 9 0 0

a

(2.6) (0) (5.3) (65.8) (0) (0) (0) (0) (0) (23.7) (0) (0)

Sum (%) 1 21 28 37 23 6 10 3 0 36 5 2

(2.6) (55.3) (73.7) (97.4) (60.5) (15.8) (26.3) (7.9) (0) (94.8) (13.2) (5.3)

Accompanied by granulomatous reaction. Predominantly centrilobular. c Centrilobular. d Small rather isomorphic vacuoles. e Involves all acinar zones in 27 cases (71.1%), scattered (without any predilection for a specific zone) in 7 cases (18.4%), predominantly involves zones 2 and 3 in 2 cases (5.3%). f Involvement of less than 5% of lobular area. b

192

S. Saleki et al. / Forensic Science International 166 (2007) 190–193

Fig. 1. Severe sinusoidal congestion with clumping of red blood cells (H&E).

Fig. 4. Sinusoidal clusters of polymorphonuclear leukocytes and mild macrovesicular steatosis (H&E).

Fig. 2. Fine vacuolization of hepatocytes (H&E). Fig. 5. Mild macrovesicular steatosis of acinar zone 3 (H&E).

Fig. 3. Hepatocyte nuclear fragmentation (H&E).

(Fig. 4), macrovesicular steatosis (Fig. 5), and subcapsular hemorrhage. The major histopathologic findings in other studies included centrilobular necrosis, congestion, vacuolar degeneration of hepatocytes, dilated central vein and sinusoids filled with proteinaceous fluid, and nuclear fragmentations [9,11]. The mechanism or nature of the above-mentioned histopathologic changes is not fully understood. However, it seems to be due to the inhibition of cytochrome C