Hairy matters: MtDNA quantity and sequence variation along and among human head hairs

Forensic Science International: Genetics 25 (2016) 1–9

Contents lists available at ScienceDirect

Forensic Science International: Genetics journal homepage: www.elsevier.com/locate/fsig

Hairy matters: MtDNA quantity and sequence variation along and among human head hairs Stijn Desmytera,1,* , Martin Bodnerb,1, Gabriela Huberb , Sophie Dognauxa , Cordula Bergerb , Fabrice Noëla , Walther Parsonb,c,** a b c

National Institute of Criminalistics and Criminology, Brussels, Belgium Institute of Legal Medicine, Medical University of Innsbruck, Innsbruck, Austria Forensic Science Program, The Pennsylvania State University, PA, USA

A R T I C L E I N F O

Article history: Received 29 May 2016 Received in revised form 20 July 2016 Accepted 23 July 2016 Available online 25 July 2016 Keywords: mtDNA heteroplasmy Intra-hair variation Intra-individual variation Hotspots Forensic evidence Casework

A B S T R A C T

Hairs from the same donor have been found to differ in mtDNA sequence within and among themselves and from other tissues, which impacts interpretation of results obtained in a forensic setting. However, little is known on the magnitude of this phenomenon and published data on systematic studies are scarce. We addressed this issue by generating mtDNA control region (CR) profiles of >450 hair fragments from 21 donors by Sanger-type sequencing (STS). To mirror forensic scenarios, we compared hair haplotypes from the same donors to each other, to the corresponding buccal swab reference haplotypes and analyzed several fragments of individual hairs. We also investigated the effects of hair color, donor sex and age, mtDNA haplogroup and chemical treatment on mtDNA quantity, amplification success and variation. We observed a wide range of individual CR sequence variation. The reference haplotype was the only or most common (
75%) hair haplotype for most donors. However, in two individuals, the reference haplotype was only found in about a third of the investigated hairs, mainly due to differences at highly variable positions. Similarly, most hairs revealed the reference haplotype along their entire length, however, about a fifth of the hairs contained up to 71% of segments with deviant haplotypes, independent of the longitudinal position. Variation affected numerous positions, typically restricted to the individual hair and in most cases heteroplasmic, but also fixed (i.e. homoplasmic) substitutions were observed. While existing forensic mtDNA interpretation guidelines were found still sufficient for all comparisons to reference haplotypes, some comparisons between hairs from the same donor could yield false exclusions when those guidelines are strictly followed. This study pinpoints the special care required when interpreting mtDNA results from hair in forensic casework. ã 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Human head hairs are common crime scene evidence [1–6]. Morphological comparison may provide investigative leads but is not trivial, especially for small fragments [7–9]. Hairs can also yield nuclear (nuc)DNA profiles when hair roots in anagenic growth phase or adhering tissue are available, but to a much lesser extent from hair shafts [1,7]. NucDNA content in general varies between

* Corresponding author at: S.D., National Institute of Criminalistics and Criminology, Vilvoordsesteenweg 100, 1120, Brussels, Belgium. ** Corresponding author at: W.P., Institute of Legal Medicine, Medical University of Innsbruck, Müllerstrasse 44, 6020, Innsbruck, Austria. E-mail addresses: [email protected] (S. Desmyter), [email protected] (W. Parson). 1 These authors contributed equally. http://dx.doi.org/10.1016/j.fsigen.2016.07.012 1872-4973/ã 2016 Elsevier Ireland Ltd. All rights reserved.

individuals [10]. Most hairs found at the crime scene were shed, which is why usually only telogenic hairs or hair fragments lacking roots are available that contain only minute or no detectable nucDNA [3,5,11–14]. Still, successful STR typing has been demonstrated also from challenging hair samples and specific DNA screening assays have been developed [10,15], however, resulting STR profiles are often incomplete [16–19]. Here, mitochondrial (mt)DNA presents a valuable genetic marker. Intact mitochondria are commonly observed in hair shafts in small numbers [7,20,21] but disappear later in development [14]. MtDNA hair casework has high success rates in specialized laboratories (cf. [22]), where
90% full or partial haplotypes were reported [2,3,5,9], even for specimens smaller than 2 mm [23]. Despite degradation [13,24], mtDNA from hairs can successfully be analyzed even in historic cases and “museomics” [9,21,25], and sophisticated strategies have been presented [24,26]. MtDNA base

2

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

changes post mortem in hairs have been discussed [27], but hairs are extremely resistant to exogenous DNA or can successfully be decontaminated [28]. In 1988, a single human head hair was proven to contain enough mtDNA to reveal an individual’s haplotype [12]. Other early mtDNA investigations determined hair as useful tissue in population studies, genetic screenings and amenable to analytical automation [29–31]. Hair mtDNA typing entered forensic casework in the mid-nineties, being introduced to courts in 1996 in Sweden and the US [3,32–36]. The significance of mtDNA in this regard is not unambiguous identification, but (non-) exclusion for samples to derive from the same individual/maternal lineage. Earlier studies described mtDNA point heteroplasmy (PHP), the presence of more than one haplotype in a single hair. Initially it was detected at disease related positions [37,38], where selection was discussed for being the driving force. Later PHP was also observed at “neutral” coding region [39] and CR positions [6]. Numerous further investigations found high mtDNA PHP rates in hairs and provided insights into particular aspects, such as familial transmission and site-specific rates, e.g., [4,9,20,40–52]. Several factors were considered to contribute to the regular occurrence and highly variable levels of mtDNA PHP in hairs. Hair follicles develop independently from different sets of stem cells. Multiple cell types, whose proportions can vary, “feed” the growing shaft and give rise to mitochondria. This might cause different ratios of mtDNA populations visible only when PHP is present. Cell contributions during initial growth, when distal hair fragments are produced, are thought to differ from the telo- or catagenic buildup of proximal parts. The extreme bottlenecks along hair shafts might cause variant ratio shifts that are better visible than in other tissues. The high energy demand during proliferation and keratinization, the long growth cycle, apoptosis, environmental effects (possibly greater for distal segments), including exposure to radiation and cosmetics, combined with the lack of a repair mechanism have been linked to a high mtDNA mutation rate in hair [4,7,22,40,41,43,44,53,54]. A major advantage of DNA analysis over morphological comparison of hairs is the possibility of a database query to yield an estimate of the probability of a match of the questioned hair haplotype with that of a randomly selected person [8]. Large and reliable databases are a prerequisite for forensic queries [55]. The limits of microscopic hair examination have been outlined in [56], where 9 of 80 associations were excluded by mtDNA. In 2009, a US National Research Council report claimed imprecision of microscopic hair analyses and reporting when considering statistics about the distribution of particular characteristics in the population and standards on the number of features “matching” hairs must agree on [57]. After reviewing 500 of its microscopic hair comparison cases prior to 2000, the FBI concluded in 2015 that
90% of trial transcripts contained statements that led to exaggeration of data significance [58]. However, this does not flaw the results of hair (mt)DNA analyses, but instead pinpoints the importance of investigations as the current one: with this study, we provide the first systematic and comprehensive insight into mtDNA variation among and along single human head hair shafts applicable to the forensic practitioner. We performed (i) a “latitudinal” hair study by comparing 25–50 hairs per capita from eleven individuals to gain better insight into intra-individual hair mtDNA variation, and (ii) a “longitudinal” hair study on 20 hair shafts from ten individuals to investigate mtDNA quantity and sequence variation along single hairs. We address shortcomings of previous investigations in mimicking forensic casework by including a reasonable number of individuals and hairs, reporting entire CR sequences from single hair shaft fragments generated by STS, the current gold standard in forensic DNA analyses, and comparing them to buccal swab reference haplotypes from the

same donors. Our aims were to investigate the degree of mtDNA variation and the existence of mutational patterns in human head hairs and to assess the validity of current forensic mtDNA interpretation guidelines that derive mainly from blood/buccal swab data. 2. Materials and methods This two-center study combines data from the National Institute of Criminalistics and Criminology, Brussels, Belgium (NICC) and the Institute of Legal Medicine, Medical University of Innsbruck, Austria (GMI). 2.1. Sample collection Plucked or shed head hairs and reference buccal swabs were collected from 21 individuals of European origin after informed consent. Donor age, sex, natural hair color and current cosmetic treatments were recorded at sampling (Table S1). The proximal centimeter of each hair including the root and possible adhering cells was removed. At NICC, 25–50 hairs per capita were collected from eight female and three male donors (23–65 years old, named P1-P11) resulting in a total of 348 hairs for testing. The first proximal 2 cm fragment of each of the 348 hairs was used for the investigation of mtDNA variation across hairs (latitudinal study). At GMI, two hairs were collected from ten women (24–36 years old, named A-J) and cut into 2 cm fragments numbered in consecutive order. This resulted in seven to 20 fragments per hair (244 in total) for a comparison of mtDNA sequence and mtDNA quantity along hairs (longitudinal study). Two cm was also the average fragment length available in 691 casework hairs [9] and is recommended in [33]. 2.2. Decontamination and DNA extraction For decontamination, the hair fragments were incubated at (a) 56  C for 1 h in 1,250 ml preferential lysis buffer A (10 mM Tris-HCl at pH 8.0, 100 mM NaCl, 10 mM EDTA, 0.5% SDS, 400 mg/ml proteinase K) and rinsed in MilliQ water at NICC, or (b) at 56  C for 2 h in 800 ml and washed thrice in 500 ml of preferential lysis buffer B (10 mM Tris-HCl at pH 8.0, 100 mM NaCl, 1 mM CaCl2, 2% SDS, 800 mg/ml proteinase K) at GMI, respectively. Total DNA was extracted (a) using the Tissue and Hair Extraction kit (for use with DNA IQ) and the DNA IQ casework sample kit on a Maxwell 16 instrument (all: Promega, Madison, WI, USA) at NICC, or (b) on a BioRobot M48 workstation (Qiagen, Hilden, Germany) after digestion in 320 ml lysis buffer B containing 20 ml 1 M DTT for 2 h or until they were completely dissolved at GMI (extraction volume: 50 ml). DNA from control buccal swabs was extracted using (a) the NucleoMag kit (Macherey-Nagel, Düren, Germany) on a KingFisher96 purification platform (Thermo, Breda, The Netherlands) at NICC, or (b) Chelex-100 (Bio-Rad, Hercules, CA, USA) [59] at GMI. 2.3. MtDNA quantitation A 143 bp-mtDNA fragment was assessed in all longitudinal hair extracts in a modular real-time quantitative PCR assay [60] on an AB7500 Fast Real-Time PCR Instrument (AB, Oyster Point, CA, USA) at GMI to compare mtDNA content along hairs and to evaluate amenability to amplification. 2.4. MtDNA amplification Complete CRs (nps 16024-576) were amplified as single fragments from the buccal swab extracts, while the amplicon

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

number for the hair segments varied: (a) at NICC, CRs were amplified in six segments of 277–320 bp in 25 ml PCR reactions containing 10 ml unquantified DNA extract, 1 mM primers (Eurogentec, Seraing, Belgium), 1 U AmpliTaq Gold polymerase (AB), 0.8 mM dNTPs, and 1.46 mM MgCl2 in 1X GeneAmp Gold PCR Buffer. The thermocycler protocol was 95  C for 11 min, 30 cycles at 94  C for 45 s, 60  C [53  C for the single amplicon from control swab extracts, resp.] for 30 s and 72  C for 2 min, and finally 72  C for 10 min; (b) at GMI, the amplification strategy was adjusted to quantitation results (cf. [26]) and different amplicons (187– 1,476 bp) were generated per extract to cover the entire CR, keeping their total number (656) as small as possible. Amplifications were carried out in 12.5, 20 or 25 ml containing up to 89,765 mitochondrial genome equivalents (mtGE) in 1X Advantage 2 Polymerase Mix, 1X Advantage 2 PCR Buffer (both: Clontech Laboratories, Mountain View, CA, USA), 0.25 mg/ml BSA (Serva, Heidelberg, Germany), 0.8 mM dNTPs, and 200 nM primer. The thermocycler protocol was 95  C for 2 min, 38 cycles at 95  C for 15 s, 56  C for 30 s, 72  C for 30–90 s (according to amplicon size). Amplification success was evaluated on an 8% PAGE gel using 3 ml of PCR product (data not shown). Buccal swab extracts were amplified as in [61]. Table S2 lists the primers used in both centers. 2.5. MtDNA sequence data generation and interpretation Sanger-type cycle sequencing was carried out using BigDye Terminator v1.1 chemistry (AB) after treating the amplification mixes with ExoSAP-IT (USB, Cleveland, OH, USA). The sequencing products were purified using (a) the DyeEX96 kit (Qiagen) at NICC or (b) Sephadex G-50 Fine (GE Healthcare, Buckinghamshire, UK) on Multiscreen filter plates (Millipore, Bedford, MA, USA) at GMI, and separated using POP6 polymer on an ABI (a) 3130xl at NICC or (b) Prism 3100 Genetic Analyzer (both AB) at GMI. Sequencing primers are listed in Table S2. Contiguous CR sequences were assembled and aligned to the revised Cambridge Reference Sequence (rCRS) [62] using (a) SeqScape v.2.5 (AB) at NICC or (b) Sequencher v.4.9 (Gene Codes, Ann Arbor, MI, USA) at GMI. Throughout this study, variation in homopolymer C-stretches was ignored (see below). Precise base calling was ensured by double independent data analysis and validation according to highest forensic quality standards [61]. The haplotypes were assigned to haplogroups of PhyloTree [63], build 17, aided by the EMMA software package [64]. 3. Results 3.1. MtDNA quantity in hair extracts MtDNA quantity (and quality) in hairs varies considerably between individuals and depends on growth phase and time since shedding/plucking [5,11,65]. Also this study revealed high intraand inter-individual differences in mtDNA content among and along hairs (Table S3). The mean mtDNA content was 2,126  1,114 mtGE/ml (53,150  27,850 mtGE/cm) from the 20 proximal 2 cm fragments. Two earlier investigations reported lower mean mtDNA content: 384  214 mtGE/ml (9,600  5,350 mtGE/cm) were measured applying the same methods in the proximal 2 cm fragments of five hairs three weeks after shedding [26], others used a different strategy and reported 20,900  35,600 mtGE/cm for the proximal 3 cm segments of shed and 40,200  35,700 mtGE/ cm of plucked hairs [65]. In our test panel, the mean mtDNA content per fragment decreased with distance from the root to 14  12 mtGE/ml (350  300 mtGE/cm) in the three 20th fragments available. The decline in mtDNA quantity along hair shafts, reported or suspected in [26,65–67], was generally stronger in hair that had undergone cosmetic treatment. In the proximal 14 cm

3

(pooled 1st–7th fragments) available from all 20 hairs included in our quantitation assays, on average 1,005  688 mtGE/ml (25,125  17,200 mtGE/cm) were observed. This mean mtDNA content was higher in untreated hairs (n = 12; 1,253  766 mtGE/ml or 31,325  19,150 mtGE/cm) than in treated hairs (n = 8; 633  318 mtGE/ml or 15,825  7,950 mtGE/cm) and varied between 1.01 and 4.73-fold in the two hairs from the donors in this test panel. Considering the entire hair shafts, the same trend was revealed: untreated hairs contained 821  495 (20,525  12,375 mtGE/cm) and treated hairs 502  322 mtGE/ml (12,550  8,050 mtGE/cm) on average (Table S3). 3.2. MtDNA amplification and sequencing success from hair extracts In the latitudinal study, i.e. the comparison of 25–50 hairs of eleven donors, 335 of 348 proximal hair shaft fragments (96%) were successfully sequenced for the CR (Table S1). From 124 of 244 extracts (51%) in the longitudinal study, i.e. the comparison of all fragments along 20 hairs of ten donors, complete CR sequences were obtained. Sequencing success mirrored the decline in mtDNA quantity towards the hair tips and from natural to treated hairs. No results were yielded from extracts containing 170 mtGE/ml (4,250 mtGE/cm) (using 5 ml extract per reaction) (Table S3). Hair H2 (low mtDNA content) and both hairs of individual B (highlighted tips) were exceptions with low success rate even though untreated. A negative effect of cosmetic treatment on amplification success has been described [68], but was not found in [32,67]. Donor age (as in [19]) and hair color/pigmentation/ melanin content (as in [5,69]) did not seem to have an influence in our samples, contrary to [9,70]. Hair texture and thickness, that could impact mtDNA yield, were not assessed in our samples. Diverging success rates between studies might be caused by individual variation, different times since shedding/plucking and co-extraction of inhibiting melanin in some protocols (cf. [68]). 3.3. MtDNA sequence variation among hairs of the same donor: results from the latitudinal study In the latitudinal study, we analyzed 25–50 proximal 2 cm hair shaft fragments of eleven individuals, in total 335 fragments. While for two donors all haplotypes were identical, several differences were detected in hair fragments of the remaining nine individuals. Three to nine mostly unique hair CR haplotypes were found in each of these nine donors. The hair segments contained up to two differences (including PHP) compared to the corresponding reference/majority haplotype, cumulating into two to nine variable nps (up to six partial or five full substitutions) per individual. For seven of the donors showing variation, the reference haplotype was the most common one (76–90%) in the analyzed hairs. Proportions were remarkably lower in two individuals: P10, whose reference haplotype carried np T195Y and was found in 36% of the hairs, exhibited np T195 and no other difference in the most common hair haplotype (48%); P11 carried np T16093C in the reference and 32% of the hairs, while the otherwise identical majority hair fragment haplotype (60%) carried np T16093Y. Disregarding these hotspots, the mtDNA variation rates in P10 (12%) and P11 (8%) dropped to the lower bound of that of the other individuals. Altogether, CR point variation in the 335 hair fragments was observed at 33 different positions as nine homoand 73 heteroplasmic events, as expected mainly transitions (cf. [9,71]). Nps 195, 16093 and 16183 were top-ranking with 18, 17 and five observations, respectively. Variation at the latter two positions was restricted to single individuals, while that at np 195 was more widespread. The remaining substitutions occurred up to three times in our test panel, but were mostly unique. None was phylogenetically informative within the individual mtDNA lineage

4

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

Table 1 Summary of differences observed between mtDNA CR sequences from reference samples and hairs of the same eleven donors in the latitudinal study. A total of 73 heteroplasmic and nine homoplasmic differences was found at 33 positions when comparing 25–50 proximal 2 cm hair fragments (n = 335). Nps

Variationa

Heteroplasmic occurrences

Homoplasmic occurrences

64 152 189 195 215 217 241 251 411 467 541 16092 16093 16145 16169 16183 16189 16221 16239 16250 16258 16265 16270 16284 16291 16294 16311 16317 16320 16354 16390 16391 16519

C/T T/C A/G T/C A/G T/C A/G G/A C/G C/T C/T T/C T/C G/A C/T A/C T/C C/T C/T C/T A/G A/G C/T A/G C/T C/T T/C A/T C/T C/T G/A G/A T/C

2 1 1 17 2 1 1 2|3d 1 1 1 3 17 0 1 4 1 0|1d 1 1 1 1 2 1 2 1 0 0 1 1 1 3 1

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

a b c d

Individualsb P4 P2 P8 P6 (1), P7 (2), P10c (15) P3, P6, P7 P4 P10 P5, P10, Dd P8 P2 P7 P6 P11 P7 P2, P8 P5 P7 P7,Gd P8 P2 P2 P10 P6 P2 P8 P4 P7 P3 P8 P4 P5, P7 P3 (1), P11 (2) P7

The first base is the rCRS variant. Numbers in brackets indicate observations per donor where necessary. Homoplasmic states in donor P10 with heteroplasmic reference were counted as PHP. Additional observations made in first proximal 2 cm hair fragments in the longitudinal hair study (Table 2).

[63], which could impact the selection of database subsets for forensic queries after phylogeographic considerations [72] (Table 1, Table S1). 3.4. MtDNA sequence variation along single hair shafts: results from the longitudinal study Also the analysis of two hairs from ten donors in their entire length, in total comprising 124 2 cm fragments, revealed CR haplotypes differing from each other and the corresponding reference. Variation was not equally distributed: while seven individuals revealed no difference, two to eight events were observed in each of the remaining three donors even though only one hair of donor D contained differences and for H, only one hair could be sequenced. All observed differences (a total of 16 at seven Table 2 Summary of differences observed between mtDNA CR sequences from reference samples and along hairs of the same ten donors in the longitudinal study. A total of 16 heteroplasmic differences was found at seven positions when comparing all 2 cm fragments along 19 hairs (n = 124). PHP

Heteroplasmic/sequenced hair fragments

242Y 251R 16183R 16221Y 16318R 16296Y 16439M

1/9 3/7 3/7 5/7 2/7 1/7 1/9

Hairs H1 D1 G1 G2 D1 D1 H1

CR positions) were heteroplasmic and one was a transversion at 16439. They were restricted to single hairs and none was phylogenetically informative [63] (Table 2, Table S1). The seven nps were found heteroplasmic in 22–71% of the sequenced fragments per hair in proximal, intermediate and distal parts, as well as recurrent along a hair with intervening homoplasmic fragments. The reference haplotype was the most common one found in all hairs but D1 and G2, and in all individuals. Only in individual G, it was not the majority profile but found in only 43% of hair fragments (Table 2, Table S1). 3.5. Influence of donor characteristics on mtDNA variation Taking the results from the longitudinal and the latitudinal part of this study together, six of the ten (60%) donors with untreated hair showed mtDNA variation in this tissue, as did six of 11 (55%) donors with cosmetically treated hair, ten of 18 (56%) female, two of three (67%) male donors, nine of 15 (60%) donors with dark (brown) and three of six (50%) donors with light (blond) hair. In our panel of 21 individuals, no association between these characteristics and the occurrence of mtDNA variation was revealed in x2 tests (data not shown). The age of donors that exhibited hair mtDNA variation (mean: 36.1 11.3, median 34 years, n = 12) was almost identical to the age of those that showed only invariant hair (mean: 35.7  11.5, median 33 years, n = 9). All mtDNA haplogroups found more than once in our panel (viz. K, R0 including H, T, and U) occurred in both donor groups (Table S1). Our findings confirm earlier studies that found no increase in hair mtDNA heteroplasmy

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

with donor age ([43,73], contrary to [44]), sex [43,73], mtDNA haplogroup [73], cosmetic treatment [43] or hair color ([7], contrary to [43]). 4. Discussion We compared nature and frequency of mtDNA variation in head hair found in this study to published records. Clearly, observations depend on the number of individuals (particularly those with heteroplasmic reference sequences) and hairs included, the mitogenome segments analyzed and the methodology applied. Variation was considered as deviation from the reference haplotype as in this study, or from other hairs, or as occurrence of heteroplasmy (with calling criteria differing between authors). In some studies, no reference sample was included. All this and a partial lack of information complicated and might bias comparisons (cf. [43]). 4.1. The latitudinal prevalence of mtDNA CR sequence variation among hairs of one donor The proportion of the eleven individuals tested in our latitudinal study that exhibited variation in their 25–50 proximal hair fragment CR haplotypes (n = 335) was 82%, compared to 38% in previous studies (Reynolds & Calloway in [20]; [42,43,46– 48,73]). When variation was present, the proportion of hair shafts per individual different to the reference haplotype was 10–68% (and 10–52% compared to the most frequent haplotype) in this study and 2–90% in earlier observations [6,20,43,49]. Some studies reported only average variant proportions over all hairs included. In our panel, 24% of all hairs were affected. Earlier studies containing several donors found proportions of 5–19% [2,9,43– 47,49,73,74] or no variation at all [19,32,33,42,67] (Table S1, Table S4). In a collaborative exercise with diverse analytical conditions, 75% of 52 hairs showed more than one haplotype [40]. 4.2. The longitudinal prevalence of mtDNA CR sequence variation along single hair shafts A single human head hair shaft can grow >1 m [19] and may yield several samples for analyses. Despite high forensic relevance, mtDNA sequence variation along hair shafts has scarcely been studied. Spontaneous investigations reported varying PHP nucleotide proportions in adjacent hair fragments from individuals with homo- [51] and heteroplasmic [4,40,66] reference haplotypes, base shifts ([51], Reynolds & Calloway in [20]), as well as PHP invariant along the shaft [6,40] and homoplasmic fragments intervening heteroplasmic ones [40]. This first systematic longitudinal examination of hair shafts performed here confirms earlier findings. Likely caused by extreme bottlenecks in hair development, fragments from one hair shaft revealed different mtDNA sequences, as if deriving from independent shafts (covered in the latitudinal part of this study). Hair roots are not considered here. They were found to differ in mtDNA sequence from their shafts in [45]; others revealed similar PHP proportions [44,52]. 4.3. Synopsis on the mtDNA positions affected by variation in hairs Only a large and systematic study might reveal particular mtDNA variation patterns in hairs. We therefore combined the results from the two parts of this study comprising mtDNA sequences of 459 hair fragments from 21 individuals (Tables 1,2). Mutational events occurring in multiple fragments of a hair shaft in the longitudinal study were counted only once for reasons of comparability with latitudinal studies. Variation was observed both at rarely or not yet described individual nps and at reported

5

hotspots in single donors in our panel (cf. [43,49]). The top-ranking nps 195, 16093 and 16183 illustrate evolutionary mechanisms of mtDNA heteroplasmy: (i) the high variation at np 195 in hair is explained by a single individual who showed this heteroplasmy already in the reference, a phenomenon reported at several nps [40,41,46,52]. Notwithstanding, np 195 was also found highly variable in individuals with homoplasmic blood in [43] using an array technique; (ii) np 16093 was the first reported [6] and the most frequent hair CR PHP in many studies [2,9,40,43,44,46– 48,51,73]. Also in our study, its top position is related to high variation rates in hairs of individuals carrying np T16093C in their reference tissue. In the latter, the transition was found both variable [2] and stable ([4], this study); while information is ambiguous in [40]. Also an individual carrying np T16093 in the reference showed variation in hairs [43]. Further nps exhibiting individual variation in hair fragments of single donors, without apparent PHP in the reference haplotype have been observed by an array technique [43], tentatively by STS [40], and using hair roots [41]. Vice versa, invariant hair segments despite np T16093C in the reference have been reported ([4], this study). It can be suspected that in many cases underlying PHP at such hypervariable nps would be found applying more sensitive techniques (see below); (iii) the third-ranking variation was the (partial) transversion A16183C. The accompanying np T16189C joined adjacent polycytosine tracts and may have caused 30 -shortening of the upstream polyadenine tract at np 16183. The co-occurrence of T16189C and a shortened A-stretch has been reported in mother-child transmissions [75] and is a phylogenetic marker [73]. Two further individuals in our test panel revealed invariant nps A16182C A16183C T16189C also in the reference. In addition, partial transitions at np 16183 occurred in one hair. The remaining sequence variation did not comprise nps occurring in more than three of the 354 individual hair shafts. Most were observed in single or two of the 21 individuals, merely nps 215 and 251 were found in three donors each in our panel (Table 1, Table 2). 4.4. The mtDNA variation in the light of previous forensic studies Our results have shown that mtDNA variation in hairs is highly impacted by individual mtDNA characteristics of the investigated donors. To diminish this bias and possibly reveal universal patterns, we combined our findings with two large-scale single hair studies that comprised 691 casework hairs [9] and 480 hairs from three donors [49], respectively. The earlier mentioned limitations in comparability may also apply here. We restricted the comparison to STS data and the greatest common mtDNA segment (nps 16024–16173 16209–16400 30–289). The three studies together recorded 172 events of mtDNA sequence variation at 73 of 602 nps in 1,630 hairs. Most frequently affected were nps 16093, 195 and 16291, observed 31, 19 and 12 times, respectively. Also, these hotspots were not equally represented in the three studies: np 16093, first in [9] and second this study, was unobserved in [49]; np 195 ranked first in this study, but was only observed once elsewhere [9]; np 16291 was one of only two nps with >1 observation in [49], not found in [9] and only twice in this study. The heterogeneously occurring top ranking nps are hence again explained by particular individuals in the test panels. Of the 30 nps in the restricted range that we report variable in our panel, 17 (57%) (each with 3 observations) were not observed in the two other studies, as were 35 of 46 nps (76%) (each with 3 observations) of [9] and 8 of 13 nps (62%) (singletons) of [49]. Only three nps, viz. 152, 189 and 16390, were found in each of the three studies (together four, five and four times, respectively), notably not nps 215 and 251, that we had tentatively considered hair mtDNA hotspots from our data. Np 215 ranked second in [9] and

6

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

fourth in the combined data, but was not observed in [49], while np 251 was only observed once [49] outside our study. Several nps throughout the combined ranking also occurred in further smaller, often HVS-I restricted hair studies, but most reports remained unique (Table S5). In summary, mtDNA variation in human head hairs appears to be highly individual even when paralleling our results with two earlier large single hair studies, albeit on Japanese [49] and US [9] populations, and several smaller investigations. A seminal forensic study combined CR data from 5,015 blood/saliva reference samples and found 319 instances of PHP at 114 nps [76]. Thirty-six nps were also found variable in the combined hair panel. Eight of these were hotspots in blood/saliva (
7 observations): the nps ranking first (16093), second (195), fourth (215) and sixth (152) in the combined hair panel, and four singletons. Np 16291, ranking third in the hair studies, was only found twice in blood/saliva [76], while 37 of 73 nps (51%) reported in the hair panel were not observed at all. The three nps present in all three large hair panels were also observed in blood/saliva [76] (Table S5). We earlier investigated CR variation in 95 deceased individuals [48]. Five hair shafts per corpse were processed together in STS to mirror other tissues included, therefore the results are not fully comparable to the single hair studies. In hairs, 39 PHPs at 27 nps were observed. Twenty-two of the latter lay in the greatest common range applied above, 14 thereof were also observed in the combined panel and additional three in other smaller hair studies. The only hair hotspot in [48] with more than three observations was np 16093, the top ranking np in the hair synopsis. Also the nps ranking second and third were reported. Notably, nps 215 (ranking fourth), 189 (ranking fifth), 251, and additional 56 of 73 nps (77%) were not observed in [48]. Thirty-six additional hair CR nps were reported variable in- and outside the overlapping mtDNA range described above ([9,40,41,43,44,46–48,51,66,73,74], this study). About half of them were also observed in blood/saliva [76] (Table S5). 4.5. Implications for forensic casework When the mtDNA haplotype generated from a hair shaft collected at a crime scene matches that of a suspect/victim or a database entry, this “non-exclusion” can connect events and might lead to donor identification. The pronounced mtDNA variation pinpointed by our results raises caveats for forensic interpretation: how different can hairs from one donor be? The handling and interpretation of heteroplasmy and base shifts have been a matter of debate in the forensic community not only in hairs (cf. [46,53]). Up to three PHPs have been reported in single hair fragments [4,39,40,44,74] and reference tissues [76,77]. Up to four PHPs were found in complete mitogenomes from single hairs [50]. High numbers of PHPs in hairs have been interpreted as mixtures/ contamination [4,9,40,78,79] or technical limitations [74]. Also single homoplasmic differences between hair fragments or to their reference haplotype have been reported earlier [6,40,43,45,46,49,52]. Unequivocally, exclusion to derive from the same donor/ maternal lineage is defined as a situation where two mtDNA haplotypes differ at two or more nps. Situations with a single difference are considered inconclusive in forensic interpretation guidelines. It has been noted that one or two differences are not straightforward in interpretation. Heteroplasmy does not contribute to an exclusion scenario [71,72,80]. According to the guidelines, none of the 459 hair fragments analyzed in this study excluded its reference haplotype, as none of the comparisons yielded more than a single homoplasmic CR difference. However, when comparing hair segments from the same individual to each other, we found in two individuals that two hair fragments yielded up to two homoplasmic CR differences (P3, P7; Table S1). Such two hair pairs

could be falsely excluded from deriving from the same donor/ maternal lineage in the absence of further information. Hairs from one individual differing mutually at two nps have been observed before, but in much smaller proportions [40,45]. The event that two samples within the same case just differ by one np has been considered very rare from a smaller database [2]. The probability of finding two hairs from one donor that show mutually two homoplasmic differences would be 0.85% (5/49 * 4/48) in donor P7, where five of 49 (10%) hairs show one homoplasmic difference to the reference haplotype each, and 0.25% (2/29 * 1/28) in donor P3, where two of 29 (7%) hairs meet this criterion. It does not appear unlikely that a single hair fragment could display two or more apparently homoplasmic differences, but probably with increasing rarity [40]. Then, the current guidelines would not suffice even when comparing hairs to reference haplotypes. Any revisions (cf. [9]) to avoid false exclusions in such cases and in comparisons of evidence hairs (possibly from multiple crime scenes) would need to be based on extended sampling. Our results do not allow general estimates of homoplasmic mtDNA variation rates. The analysis of multiple hairs of one donor may confirm or reveal underlying heteroplasmy and additional variation (cf. [9,40], this study). A composite haplotype from “averaging” separately analyzed hair fragments (cf. [48]) will in most cases resemble the reference tissue haplotype (possibly contained in a database), as our results perfectly illustrate. The “consensus” strategy minimizes the risk of reporting false exclusion and contamination, especially when analyzing possibly degraded hairs, but can usually only be applied for known samples [2,7,20,40,41,44,45,52,66,81,82]. This approach however, would be laborious and expensive. Alternatively, pooling of multiple hairs into one extraction assay could serve as additional source of information. More research would be necessary to evaluate the significance of such an analysis: while the predominant variants will be detected, the effect of variation present in small proportions as well as mutation rate information might be lost. Still, a single hair’s haplotype should be taken as valid [20]  the chance of false exclusion of its reference haplotype may be very low according to our results. Depending on the np [51], PHPs encountered in several hair fragments likely constitute “pervasive” heteroplasmy possibly present also in other tissues [20,48]. MtDNA PHPs identical by descent may increase the strength of evidence confirmatory of a relationship [20,45,71,72]. When heteroplasmy is encountered in a query sample, there might be no sample in an mtDNA database sharing it. The search algorithm applied is crucial for a reliable outcome of database queries: in case the PHP was omitted (as it anyhow would not count as forensic difference), the affected np would be assumed to be in the reference sequence state, rendering the other variant a full difference. As the direction of mutation is not known, a connected database entry might carry any of the two homoplasmic nucleotides (cf. [45]), just as we found both losses and gains of differences to the reference sequence in hairs of the same donor in this study (Table S1). The pattern match algorithm implemented into the European DNA Profiling Group MtDNA Population Database (EMPOP) [55] takes the biological nature of heteroplasmy into account and offers a solution where it “does not result in exclusion to either of the contributing bases” [72]. 4.6. Notes on low level mtDNA variation and length heteroplasmy Investigations that applied multiple techniques and collaborative exercises where “identical” hair samples resulted in heteroand either homoplasmic base calls exemplify that observed homoplasmic differences might often constitute PHP below detection level [2,4,6,7,40,41,44,45,49,51,52,82,83]. PHP revelation and proportions depend on method, instrumentation and software. Peak height/area quantitation in STS is not reliable and a

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

fixed detection threshold value is not always feasible [2,4,20,40,72,84]. Minor nucleotide proportion detection thresholds of
10% [33,42] have been determined for STS; with methodological evolution, now <10% are considered realistic [48,85] without sub-cloning suspected mtDNA variants. Recommendations encourage laboratories to develop guidelines according to their experience and evaluate every heteroplasmy per se [71,72,80]. From rarely applied alternative sequencing methods that are generally more sensitive [20,43,49,86], although possibly not always [45,47], it can be suspected that PHP is more frequent than visible from STS. With the emerging massively parallel sequencing (MPS) methods, lower PHP detection thresholds can be expected due to higher coverage (cf. [48,87,88]). Hair shaft mtDNA MPS has recently been shown in a forensic environment [50]. Homopolymer stretch length variation (except at np 16183) was beyond the scope of this study but has been investigated (in hair) elsewhere, e.g., [42,53,89,90]. The dominant length variants are listed in Table S1. Insights from this study will contribute to a follow-up investigation on the scoring of LHP, as encouraged by current guidelines that leave “representation and reporting of LHP in forensic casework” up to the individual laboratories [72]. 5. Conclusions This first comprehensive and systematic investigation of the complete CR of human head hairs by STS did not reveal hairspecific variation hotspots or patterns that hold true also in other large studies. Our results confirm that hair mtDNA variation in the CR is extremely heterogeneous concerning rates and nps affected, mostly heteroplasmic and consists mainly of transitions (Tables 1,2). The variable mtDNA positions seemed to constitute two groups. The vast majority was only found in few individuals in single or two investigations. They appeared to be of probably somatic origin and may depend on the cells in individual hair bulbs and sequence surroundings (cf. [2,9,40,47,49,74,76,78]). In parallel, a few hypervariable mtDNA sites occurred. They were neither omnipresent among donors or studies nor hair-specific but frequently observed also in other tissues. These hotspots, reflected in the overlapping mutational patterns of blood/saliva and hairs, included nps 16093 and 152, to a smaller extent also nps 189, 195, 215, and 16291 (Table S5). We found that intra-individual mtDNA discordance in hair shafts is not exceptional as postulated [19], but still does not violate casework recommendations. According to our STS results, the current interpretation guidelines are appropriate for the vast majority of cases, but not always when hair mtDNA haplotypes of one donor are mutually compared. Pattern database searches may assist forensic examination (cf. [40]). Future studies will need to answer the remaining questions on (hair) mtDNA variation by specifically investigating the impact of certain donor characteristics, such as mtDNA haplogroup and specific population backgrounds. It has been desired to empirically estimate tissueand site-specific mutation rates to assign numerical values to given differences or matches and consider likelihood measures rather than fixed numbers of differences. For now, insufficient data are available not only from hair and all variation is treated equal [4,20,46,51]. Extended insights will likely influence our perspectives on evolution, dispersal and interpretation of mtDNA heteroplasmy. Acknowledgements The authors are grateful to the individuals that donated their hairs for research and wish to thank Sylvie Comblez, Isabel Verbeke, Valérie Decroyer, and Greet De Cock (all: NICC, Brussels, Belgium) for technical help.

7

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. fsigen.2016.07.012. References [1] C.A. Linch, S.L. Smith, J.A. Prahlow, Evaluation of the human hair root for DNA typing subsequent to microscopic comparison, J. Forensic Sci. 43 (1998) 305– 314. [2] T. Melton, K. Nelson, Forensic mitochondrial DNA analysis: two years of commercial casework experience in the United States, Croat. Med. J. 42 (2001) 298–303. [3] M. Allen, A.S. Engstrom, S. Meyers, O. Handt, T. Saldeen, A. von Haeseler, S. Pääbo, U. Gyllensten, Mitochondrial DNA sequencing of shed hairs and saliva on robbery caps: sensitivity and matching probabilities, J. Forensic Sci. 43 (1998) 453–464. [4] A. Alonso, A. Salas, C. Albarran, E. Arroyo, A. Castro, M. Crespillo, A.M. di Lonardo, M.V. Lareu, C.L. Cubria, M.L. Soto, J.A. Lorente, M.M. Semper, A. Palacio, M. Paredes, L. Pereira, A.P. Lezaun, J.P. Brito, A. Sala, M.C. Vide, M. Whittle, J.J. Yunis, J. Gomez, Results of the 1999–2000 collaborative exercise and proficiency testing program on mitochondrial DNA of the GEP-ISFG: an inter-laboratory study of the observed variability in the heteroplasmy level of hair from the same donor, Forensic Sci. Int. 125 (2002) 1–7. [5] M. Nilsson, S. Norlin, M. Allen, Sequencing of mtDNA in shed hairs: a retrospective analysis of material from forensic cases and a pre-screening method, Open Forensic Sci. J. 5 (2012) 13–22. [6] K.M. Sullivan, R. Alliston-Greiner, F.I.A. Archampong, R. Piercy, G. Tully, P. Gill, C. Lloyd-Davies, A single difference in mtDNA control region sequence observed between hair shaft and reference samples from a single donor, Proceedings from the 7th International Symposium on Human Identification 1996, Promega Corporation, Madison, WI, USA, 1997, pp. 126–130. [7] C.A. Linch, D.A. Whiting, M.M. Holland, Human hair histogenesis for the mitochondrial DNA forensic scientist, J. Forensic Sci. 46 (2001) 844–853. [8] W.F. Rowe, The current status of microscopical hair comparisons, Sci. World J. 1 (2001) 868–878. [9] T. Melton, G. Dimick, B. Higgins, L. Lindstrom, K. Nelson, Forensic mitochondrial DNA analysis of 691 casework hairs, J. Forensic Sci. 50 (2005) 73–80. [10] S. Szabo, K. Jaeger, H. Fischer, E. Tschachler, W. Parson, L. Eckhart, In situ labeling of DNA reveals interindividual variation in nuclear DNA breakdown in hair and may be useful to predict success of forensic genotyping of hair, Int. J. Legal Med. 126 (2012) 63–70. [11] C.F. Bengtsson, M.E. Olsen, L.O. Brandt, M.F. Bertelsen, E. Willerslev, D.J. Tobin, A.S. Wilson, M.T. Gilbert, DNA from keratinous tissue. Part I: hair and nail, Ann. Anat. 194 (2012) 17–25. [12] R. Higuchi, C.H. von Beroldingen, G.F. Sensabaugh, H.A. Erlich, DNA typing from single hairs, Nature 332 (1988) 543–546. [13] E. Jehaes, K. Toprak, N. Vanderheyden, H. Pfeiffer, J.J. Cassiman, B. Brinkmann, R. Decorte, Pitfalls in the analysis of mitochondrial DNA from ancient specimens and the consequences for forensic DNA analysis: the historical case of the putative heart of Louis XVII, Int. J. Legal Med. 115 (2001) 135–141. [14] C.A. Linch, Degeneration of nuclei and mitochondria in human hairs, J. Forensic Sci. 54 (2009) 346–349. [15] L. Bourguignon, B. Hoste, T. Boonen, K. Vits, F. Hubrecht, A fluorescent microscopy-screening test for efficient STR-typing of telogen hair roots, Forensic Sci. Int.-Genet. 3 (2008) 27–31. [16] D. McNevin, L. Wilson-Wilde, J. Robertson, J. Kyd, C. Lennard, Short tandem repeat (STR) genotyping of keratinised hair. Part 2. An optimised genomic DNA extraction procedure reveals donor dependence of STR profiles, Forensic Sci. Int. 153 (2005) 247–259. [17] S. Amory, C. Keyser, E. Crubezy, B. Ludes, STR typing of ancient DNA extracted from hair shafts of Siberian mummies, Forensic Sci. Int. 166 (2007) 218–229. [18] K. Muller, R. Klein, E. Miltner, P. Wiegand, Improved STR typing of telogen hair root and hair shaft DNA, Electrophoresis 28 (2007) 2835–2842. [19] H. Pfeiffer, J. Huhne, C. Ortmann, K. Waterkamp, B. Brinkmann, Mitochondrial DNA typing from human axillary, pubic and head hair shafts–success rates and sequence comparisons, Int. J. Legal Med. 112 (1999) 287–290. [20] T. Melton, Mitochondrial DNA heteroplasmy, Forensic Sci. Rev. 16 (2004) 1–20. [21] H. Wolinsky, History in a single hair, EMBO Rep. 11 (2010) 427–430. [22] L. Prieto, M. Montesino, A. Salas, A. Alonso, C. Albarrán, S. Álvarez, M. Crespillo, A.M. Di Lonardo, C. Doutremepuich, I. Fernández-Fernández, A.G. de la Vega, L. Gusmão, C.M. López, M. López-Soto, J.A. Lorente, M. Malaghini, C.A. Martı’nez, N.M. Modesti, A.M. Palacio, M. Paredes, S.D.J. Pena, A. Pérez-Lezaun, J.J. Pestano, J. Puente, A. Sala, M.C. Vide, M.R. Whittle, J.J. Yunis, J. Gómez, The 2000–2001 GEP–ISFG collaborative exercise on mtDNA: assessing the cause of unsuccessful mtDNA PCR amplification of hair shaft samples, Forensic Sci. Int. 134 (2003) 46–53. [23] T. Melton, G. Dimick, B. Higgins, M. Yon, C. Holland, Mitochondrial DNA analysis of 114 hairs measuring less than 1 cm from a 19-year-old homicide, Invest. Genet. 3 (2012) 12.

8

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9

[24] S. Köhnemann, P. Pennekamp, P.F. Schmidt, H. Pfeiffer, qPCR and mtDNA SNP analysis of experimentally degraded hair samples and its application in forensic casework, Int. J. Legal Med. 124 (2010) 337–342. [25] M.T. Gilbert, A.S. Wilson, M. Bunce, A.J. Hansen, E. Willerslev, B. Shapiro, T.F. Higham, M.P. Richards, T.C. O'Connell, D.J. Tobin, R.C. Janaway, A. Cooper, Ancient mitochondrial DNA from hair, Curr. Biol. 14 (2004) R463–464. [26] C. Berger, W. Parson, Mini-midi-mito: adapting the amplification and sequencing strategy of mtDNA to the degradation state of crime scene samples, Forensic Sci. Int. Genet. 3 (2009) 149–153. [27] M.T. Gilbert, R.C. Janaway, D.J. Tobin, A. Cooper, A.S. Wilson, Histological correlates of post mortem mitochondrial DNA damage in degraded hair, Forensic Sci. Int. 156 (2006) 201–207. [28] M.T. Gilbert, L. Menez, R.C. Janaway, D.J. Tobin, A. Cooper, A.S. Wilson, Resistance of degraded hair shafts to contaminant DNA, Forensic Sci. Int. 156 (2006) 208–212. [29] L. Vigilant, R. Pennington, H. Harpending, T.D. Kocher, A.C. Wilson, Mitochondrial-DNA sequences in single hairs from a southern african population, Proc. Natl. Acad. Sci. U. S. A. 86 (1989) 9350–9354. [30] D.M. Thomson, N.N. Brown, A.E. Clague, Routine use of hair root or buccal swab specimens for pcr analysis–advantages over using blood, Clin. Chim. Acta 207 (1992) 169–174. [31] R. Hopgood, K.M. Sullivan, P. Gill, Strategies for automated sequencing of human mitochondrial-DNA directly from PCR products, Biotechniques 13 (1992) 82–92. [32] M.R. Wilson, J.A. DiZinno, D. Polanskey, J. Replogle, B. Budowle, Validation of mitochondrial DNA sequencing for forensic casework analysis, Int. J. Legal Med. 108 (1995) 68–74. [33] M.R. Wilson, D. Polanskey, J. Butler, J.A. DiZinno, J. Replogle, B. Budowle, Extraction, PCR amplification and sequencing of mitochondrial DNA from human hair shafts, Biotechniques 18 (1995) 662–669. [34] R. Decorte, E. Jehaes, F.-X. Xiao, J.-J. Cassiman, Genetic analysis of single hair shafts by automated sequence analysis of the mitochondrial D-loop region, Advances in Forensic Haemogenetics, 6, Springer, Berlin, 1995, pp. 17–19. [35] 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 (1998) 124– 132. [36] C.L. Davis, Mitochondrial DNA: State of Tennessee v. Paul Ware, Profiles in DNA 1 (1998) 6–7. [37] N. Kotsimbos, M.J.B. Jean-Francois, M. Huizing, R.M.I. Kapsa, P. Lertrit, N.C. Siregar, S. Marzuki, C. Sue, E. Byrne, Rapid and noninvasive screening of patients with mitochondrial myopathy, Hum. Mutat. 4 (1994) 132–135. [38] H.L. Hao, E. Bonilla, G. Manfredi, S. Dimauro, C.T. Moraes, Segregation patterns of a novel mutation in the mitochondrial transfer-RNA glutamic-acid gene associated with myopathy and diabetes-mellitus, Am. J. Hum. Genet. 56 (1995) 1017–1025. [39] D. Comas, S. Pääbo, J. Bertranpetit, Heteroplasmy in the control region of human mitochondrial DNA, Genome Res. 5 (1995) 89–90. [40] G. Tully, S.M. Barritt, K. Bender, E. Brignon, C. Capelli, N. Dimo-Simonin, C. Eichmann, C.M. Ernst, C. Lambert, M.V. Lareu, B. Ludes, B. Mevag, W. Parson, H. Pfeiffer, A. Salas, P.M. Schneider, E. Staalstrom, Results of a collaborative study of the EDNAP group regarding mitochondrial DNA heteroplasmy and segregation in hair shafts, Forensic Sci. Int. 140 (2004) 1–11. [41] K.E. Bendall, V.A. Macaulay, B.C. Sykes, Variable levels of a heteroplasmic point mutation in individual hair roots, Am. J. Hum. Genet. 61 (1997) 1303–1308. [42] G.G. Paneto, L.V. Longo, J.A. Martins, M.A. de Camargo, J.C. Costa, A.C. de Mello, B. Chen, R.N. Oliveira, M.H. Hirata, R.M. Cicarelli, Heteroplasmy in hair: study of mitochondrial DNA third hypervariable region in hair and blood samples, J. Forensic Sci. 55 (2010) 715–718. [43] K.A. Roberts, C. Calloway, Characterization of mitochondrial DNA sequence heteroplasmy in blood tissue and hair as a function of hair morphology, J. Forensic Sci. 56 (2011) 46–60. [44] Q. Wang, R.G. Boles, Individual human hair mitochondrial DNA control region heteroplasmy proportions in mothers and children, Mitochondrion 6 (2006) 37–42. [45] K. Sekiguchi, K. Kasai, B.C. Levin, Inter- and intragenerational transmission of a human mitochondrial DNA heteroplasmy among 13 maternally-related individuals and differences between and within tissues in two family members, Mitochondrion 2 (2003) 401–414. [46] G.G. Paneto, J.A. Martins, L.V. Longo, G.A. Pereira, A. Freschi, V.L. Alvarenga, B. Chen, R.N. Oliveira, M.H. Hirata, R.M. Cicarelli, Heteroplasmy in hair: differences among hair and blood from the same individuals are still a matter of debate, Forensic Sci. Int. 173 (2007) 117–121. [47] A. Brandstätter, W. Parson, Mitochondrial DNA heteroplasmy or artefacts-a matter of the amplification strategy? Int. J. Legal Med. 117 (2003) 180–184. [48] J. Naue, S. Hörer, T. Sanger, C. Strobl, P. Hatzer-Grubwieser, W. Parson, S. LutzBonengel, Evidence for frequent and tissue-specific sequence heteroplasmy in human mitochondrial DNA, Mitochondrion 20 (2015) 82–94. [49] K. Sekiguchi, H. Sato, K. Kasai, Mitochondrial DNA heteroplasmy among hairs from single individuals, J. Forensic Sci. 49 (2004) 986–991. [50] W. Parson, G. Huber, L. Moreno, M.B. Madel, M.D. Brandhagen, S. Nagl, C. Xavier, M. Eduardoff, T.C. Callaghan, J.A. Irwin, Massively parallel sequencing of complete mitochondrial genomes from hair shaft samples, Forensic Sci. Int. Genet. 15 (2015) 8–15.

[51] A. Salas, M.V. Lareu, A. Carracedo, Heteroplasmy in mtDNA and the weight of evidence in forensic mtDNA analysis: a case report, Int. J. Legal Med. 114 (2001) 186–190. [52] M.R. Wilson, D. Polanskey, J. Replogle, J.A. DiZinno, B. Budowle, A family exhibiting heteroplasmy in the human mitochondrial DNA control region reveals both somatic mosaicism and pronounced segregation of mitotypes, Hum. Genet. 100 (1997) 167–171. [53] H. Pfeiffer, S. Lutz-Bonengel, S. Pollak, R. Fimmers, M.P. Baur, B. Brinkmann, Mitochondrial DNA control region diversity in hairs and body fluids of monozygotic triplets, Int. J. Legal Med. 118 (2004) 71–74. [54] A. Junge, B. Madea, Die Rolle der humanen mitochondrialen DNA in der Haaranalytik, in: B. Madea, F. Mußhoff (Eds.), Haaranalytik: Technik und Interpretation in Medizin und Strafrecht, Deutscher Ärzte-Verlag, Köln, 2004, pp. 364–379. [55] W. Parson, A. Dür, EMPOP-a forensic mtDNA database, Forensic Sci. Int. Genet. 1 (2007) 88–92 (http://empop.online). [56] M.M. Houck, B. Budowle, Correlation of microscopic and mitochondrial DNA hair comparisons, J. Forensic Sci. 47 (2002) 964–967. [57] National Research Council Committee on Identifying the Needs of the Forensic Sciences Community, Strengthening Forensic Science in the United States: A Path Forward, The National Academies Press, Washington, DC, 2009. [58] FBI, Press release Apr 20, 2015 last accessed on July 18, 2016 at http://www.fbi. gov/news/pressrel/press-releases/fbi-testimony-on-microscopic-hairanalysis-contained-errors-in-at-least-90-percent-of-cases-in-ongoingreview. [59] 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. [60] H. Niederstätter, S. Köchl, P. Grubwieser, M. Pavlic, M. Steinlechner, W. Parson, A modular real-time PCR concept for determining the quantity and quality of human nuclear and mitochondrial DNA, Forensic Sci. Int. Genet. 1 (2007) 29– 34. [61] W. Parson, H.J. Bandelt, Extended guidelines for mtDNA typing of population data in forensic science, Forensic Sci. Int. Genet. 1 (2007) 13–19. [62] 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 (1999) 147. [63] M. van Oven, M. Kayser, Updated comprehensive phylogenetic tree of global human mitochondrial DNA variation, Hum. Mutat. 30 (2009) E386–394. [64] A.W. Röck, A. Dür, M. van Oven, W. Parson, Concept for estimating mitochondrial DNA haplogroups using a maximum likelihood approach (EMMA), Forensic Sci. Int.-Genet. 7 (2013) 601–609. [65] H. Andreasson, M. Nilsson, B. Budowle, H. Lundberg, M. Allen, Nuclear and mitochondrial DNA quantification of various forensic materials, Forensic Sci. Int. 164 (2006) 56–64. [66] K. Sekiguchi, K. Imaizumi, H. Matsuda, N. Mizuno, K. Yoshida, H. Senju, H. Sato, K. Kasai, MtDNA sequence analysis using capillary electrophoresis and its application to the analysis of mtDNA in hair, Jpn. J. Sci. Technol. Ident. 7 (2003) 123–130. [67] J. Hühne, H. Pfeiffer, K. Waterkamp, B. Brinkmann, Mitochondrial DNA in human hair shafts-existence of intra-individual differences? Int. J. Legal Med. 112 (1999) 172–175. [68] L. Pötsch, K. Bender, Untersuchungen zu möglichen Mechanismen einer PCRHemmung in Haarextrakten durch Melanine, Rechtsmedizin 11 (2001) 42–45. [69] C.A. Linch, J.R. Champagne, M.D. Bonnette, T. Dawson Cruz, Specific melanin content in human hairs and mitochondrial DNA typing success, Am. J. Forensic Med. Pathol. 30 (2009) 162–166. [70] K.A. Roberts, C. Calloway, Mitochondrial DNA amplification success rate as a function of hair morphology, J. Forensic Sci. 52 (2007) 40–47. [71] G. Tully, W. Bär, B. Brinkmann, A. Carracedo, P. Gill, N. Morling, W. Parson, P. Schneider, Considerations by the European DNA profiling (EDNAP) group on the working practices, nomenclature and interpretation of mitochondrial DNA profiles, Forensic Sci. Int. 124 (2001) 83–91. [72] W. Parson, L. Gusmao, D.R. Hares, J.A. Irwin, W.R. Mayr, N. Morling, E. Pokorak, M. Prinz, A. Salas, P.M. Schneider, T.J. Parsons, DNA Commission of the International Society for Forensic Genetics: revised and extended guidelines for mitochondrial DNA typing, Forensic Sci. Int. Genet. 13 (2014) 134–142. [73] M.A. de Camargo, G.G. Paneto, A.C. de Mello, J.A. Martins, W. Barcellos, R.M. Cicarelli, No relationship found between point heteroplasmy in mitochondrial DNA control region and age range, sex and haplogroup in human hairs, Mol. Biol. Rep. 38 (2011) 1219–1223. [74] T. Grzybowski, B.A. Malyarchuk, J. Czarny, D. Miscicka-Sliwka, R. Kotzbach, High levels of mitochondrial DNA heteroplasmy in single hair roots: reanalysis and revision, Electrophoresis 24 (2003) 1159–1165. [75] L. Forster, P. Forster, S.M. Gurney, M. Spencer, C. Huang, A. Rohl, B. Brinkmann, Evaluating length heteroplasmy in the human mitochondrial DNA control region, Int. J. Legal Med. 124 (2010) 133–142. [76] J.A. Irwin, J.L. Saunier, H. Niederstätter, K.M. Strouss, K.A. Sturk, T.M. Diegoli, A. Brandstätter, W. Parson, T.J. Parsons, Investigation of heteroplasmy in the human mitochondrial DNA control region: a synthesis of observations from more than 5000 global population samples, J. Mol. Evol. 68 (2009) 516–527. [77] S. Morovvati, Z. Morovvati, R. Ranjbar, Detecting rare triple heteroplasmic substitutions in the mitochondrial DNA control region: a potential concern for forensic DNA studies, Cell J. 13 (2011) 103–106.

S. Desmyter et al. / Forensic Science International: Genetics 25 (2016) 1–9 [78] C.D. Calloway, R.L. Reynolds, G.L. Herrin Jr., W.W. Anderson, The frequency of heteroplasmy in the HVII region of mtDNA differs across tissue types and increases with age, Am. J. Hum. Genet. 66 (2000) 1384–1397. [79] E.A. Graffy, D.R. Foran, A simplified method for mitochondrial DNA extraction from head hair shafts, J. Forensic Sci. 50 (2005) 1119–1122. [80] Scientific Working Group on DNA Analysis Methods (SWGDAM) Interpretation guidelines for mitochondrial DNA (mtDNA) analysis for forensic DNA testing laboratories, 2013. [81] B. Budowle, J.A. DiZinno, M.R. Wilson, Interpretation Guidelines for Mitochondrial DNA Sequencing, Proceedings from the 10th International Symposium on Human Identification 1999, Promega Corporation, Madison, WI, USA, 2000. [82] M.M. Holland, T.J. Parsons, Mitochondrial DNA sequence analysis–validation and use for forensic casework, Forensic Sci. Rev. 11 (1999) 21–50. [83] V. Kaplanova, J. Zeman, H. Hansikova, L. Cerna, H. Houst'kova, N. Misovicova, J. Houstek, Segregation pattern and biochemical effect of the G3460A mtDNA mutation in 27 members of LHON family, J. Neurol. Sci. 223 (2004) 149–155. [84] W. Parson, A. Brandstätter, A. Alonso, N. Brandt, B. Brinkmann, A. Carracedo, D. Corach, O. Froment, I. Furac, T. Grzybowski, K. Hedberg, C. Keyser-Tracqui, T. Kupiec, S. Lutz-Bonengel, B. Mevag, R. Ploski, H. Schmitter, P. Schneider, D. Syndercombe-Court, E. Sorensen, H. Thew, G. Tully, R. Scheithauer, The EDNAP mitochondrial DNA population database (EMPOP) collaborative exercises: organisation, results and perspectives, Forensic Sci. Int. 139 (2004) 215–226.

9

[85] J. Naue, T. Sanger, U. Schmidt, R. Klein, S. Lutz-Bonengel, Factors affecting the detection and quantification of mitochondrial point heteroplasmy using Sanger sequencing and SNaPshot minisequencing, Int. J. Legal Med. 125 (2011) 427–436. [86] L.A. Tully, T.J. Parsons, R.J. Steighner, M.M. Holland, M.A. Marino, V.L. Prenger, A sensitive denaturing gradient-gel electrophoresis assay reveals a high frequency of heteroplasmy in hypervariable region 1 of the human mtDNA control region, Am. J. Hum. Genet. 67 (2000) 432–443. [87] R.S. Just, J.A. Irwin, W. Parson, Mitochondrial DNA heteroplasmy in the emerging field of massively parallel sequencing, Forensic Sci. Int. Genet. 18 (2015) 131–139. [88] W. Parson, C. Strobl, G. Huber, B. Zimmermann, S.M. Gomes, L. Souto, L. Fendt, R. Delport, R. Langit, S. Wootton, R. Lagace, J. Irwin, Evaluation of next generation mtGenome sequencing using the Ion torrent personal genome machine (PGM), Forensic Sci. Int. Genet. 7 (2013) 543–549. [89] S. Lutz-Bonengel, T. Sanger, S. Pollak, R. Szibor, Different methods to determine length heteroplasmy within the mitochondrial control region, Int. J. Legal Med. 118 (2004) 274–281. [90] C. Berger, P. Hatzer-Grubwieser, C. Hohoff, W. Parson, Evaluating sequencederived mtDNA length heteroplasmy by amplicon size analysis, Forensic Sci. Int.-Gen. 5 (2011) 142–145.