Real-time PCR assays for the detection of tissue and body fluid specific mRNAs

Fang R¹, Manohar C², Shulse C¹, Brevnov M¹,Wong A¹, Petrauskene OV¹, Brzoska P¹, Furtado MR¹

¹Applied Biosystems, Foster City CA

²Lawrence Livermore National Laboratories, Livermore CA.

Identification of tissue parts and body fluids is frequently required in crime scene investigations. Conventional methods are often labor-intensive, not confirmatory and employ a diverse range of methodologies. Several forensic laboratories have pioneered the selection of specific protein or mRNA markers for identification of tissues and body fluids. Applied Biosystems has designed and tested real-time PCR based Taqman^â assays that target the detection of over 20,000 mRNAs encoded by the human genome. We have employed proprietary methods to design assays specific to a target transcript avoiding amplification of related gene transcripts. We have developed methods for extraction of both RNA and DNA from samples. We have also developed methods for pre-amplification of hundreds of targets present in a single sample preserving relative quantification information. These methods will be useful when dealing with heterogeneous mixtures.

In this study we have tested the performance of assays targeting saliva specific markers, Statherin, Histatin, PRB1, PRB2, PRB3; menstrual blood markers like mettaloproteinases; and semen specific markers like protamines. Data will be presented to demonstrate the capability to pre-amplify small amounts of RNA enabling testing for the presence of multiple mRNA species when the amount of RNA is limiting. Capability to multiplex these assays will also be presented. 

Contact: furtadmr@appliedbiosystems.com 

O-23

Determination of forensically relevant SNPs in MC1R gene

Branicki W¹, Kupiec T¹, Wolańska-Nowak P¹, Brudnik U²

¹ Institute of Forensic Research, Cracow, Poland

² Collegium Medicum of the Jagiellonian University, Cracow, Poland

High variation present among humans in pigmentation causes that genetic prediction of this physical trait seems attractive for forensic investigations. Genetic typing of biological traces collected at scenes of crime could be a source of valuable information about the donor’s characteristics. More than 60 genes are expected to be involved in the process of pigmentation in humans, but until present the only gene which influence for physiological variation on human pigmentation has been proved is the melanocortin 1 receptor gene (MC1R). The MC1R plays a key role in eumelanin/pheomelanin ratio in humans and hence its influence on hair and skin colour is crucial. Some allelic variants of the MC1R are significantly associated with the overproduction of pheomelanin, which is manifested with such phenotypic features as red hair or light skin. It has been suggested that analysis of MC1R variation could serve as a good indicator of the red hair phenotype. However, the postulated dosage effect of the MC1R variants on pigment phenotype disables the simple inference in the red/ non red mode. The influence of other genes on ultimate hair or skin colour makes the analysis even more complicated. Our goal was to check the variation within MC1R gene characteristic for Polish population and evaluate the usefulness of its analysis in forensic studies. A complete sequence data determined for the MC1R gene revealed, that in our population, red hair colour is mainly associated with the following variants: R151C, R160W and D294H what remains in good concordance with data for other European population samples. In our region blond-red hair phenotype is relatively common and seems mostly associated with heterozygotes or compound heterozygotes for the above alleles. Pure red hair colour can be, however, associated with homozygotes and compound heterozygotes. Hence, using simple sequence analysis more certain conclusions predicting the pure red hair colour can be drawn only for homozygous individuals, who are poorly represented in the studied population sample. Individual cases suggesting actions of other genes that could mask the influence of the MC1R variants on pigmentary status or determine a similar pigmentary effect has also been noted. Additionally a SNaPshot based assay has been developed, providing a selective analysis of the variable sites within MC1R gene, which have a significant correlation with red hair. Performed validation confirmed that the developed test enables reliable analysis of forensic specimens. We can conclude that at present the forensic usefulness of MC1R SNPs is of rather low value, but the growing data on association of particular gene variants with different phenotypic characteristics allow us to optimistically look ahead.

Contact: tkupiec@ies.krakow.pl

O-24

Hair colour in Danish families: Genetic screening of 15 SNPs in the MC1R gene by analysis of a multiplexed SBE reaction using capillary electrophoresis or MALDI-TOF MS

Mengel-Jørgensen J¹, Eiberg H², Børsting C¹, Morling N¹

¹Department of Forensic Genetics, Institute of Forensic Medicine, University of Copenhagen, Denmark.

²Institute of Medical Biochemistry and Genetics, University of Copenhagen, Denmark

Hair, eye and skin pigmentation in humans is a result of the synthesis and the deposition of melanin. The Melanocortin 1 Receptor (MC1R) is an important regulator of melanin synthesis and numerous mutations in the single-exon MC1R gene encoding MC1R have been reported. Some of these mutations affect the function of MC1R and they have been found in high frequencies in individuals with red and blond hair. A total of 15 SNPs from the MC1R gene were selected: Eight missence mutations (V60L, D84C, V92M, R142H, R151C, R160W, R163Q, D294H), two insertion mutations (179InsC, 29InsA), two silent mutations (P300P, T314T) and three SNPs near the important regulatory element, SP-1, in the MC1R promoter (rs3212359, rs3212360, rs3212361). Two PCR strategies were applied. Five short fragments covering 793 bp of the MC1R gene were amplified in a multiplex PCR to allow amplification of DNA purified from decomposed samples. Alternatively, a 1,648 bp fragment covering the entire coding region, 626 bp of the promoter and 59 bp downstream of the coding region was amplified with the purpose of determining the haplotypes of selected samples. The 15 SNPs were typed with a multiplexed single base extension reaction and detected by either capillary electrophoresis or MALDI-TOF MS. Examples of MC1R SNPs in Danish families with red haired members will be presented. Red haired individuals were typically homozygous for the mutant allele at one locus or compound heterozygous for two of the selected loci.

Contact: Jonas.Mengel@forensic.ku.dk 

O-25

Initial Study of Candidate Genes on Chromosome 2 for Relative Hand Skill

Phillips C¹, Barbaro A^1,2, Lareu MV¹, Salas A¹ and Carracedo A¹

1. Institute of Legal Medicine, University of Santiago de Compostela, Galicia, Spain

2. SIMEF, 4, Via Nicolò da Reggio, Reggio Calabria, Italy

Relative hand skill or handedness (HSR, OMIM: 139900) is a physical characteristic trait that divides people into two groups: one comprising 89-91% of individuals with a preference to use the right hand for complex manipulative tasks (typically handwriting) and the other comprising 9-11% with a left hand preference. Until recently the trait was thought to have a significant environmental component and a low heritability (1) due principally to the repeated observation of discordance for hand skill in half of left handed monozygotic twins studied, together with, often, inadequate measurement of subjects in hand skill studies. However, following the development of the random recessive, non-determinate theory for the genetic control of hand skill and other laterality traits, a robust and predictive model now exists that is consistent with the simple, mendelian inheritance of a single locus. This model implies a recessive allele frequency of ~0.48 based on an observed total of 18% discordant individuals amongst all monozygotic twin groups tested to date (2).

We have used the results of two STR based linkage analysis studies that measured hand skill as a quantitative trait (3, 4) to focus on a 3.5Mb peri-centromeric region of chromosome 2 to search for candidate genes. We aim to refine the linkage signal, initially genotyping a reduced subset of the coding SNPs in 42 genes found in the region defined by the strongest signal previously reported from a limited STR marker set. The SNP genotyping efforts required to scan such a large group of loci are considerable. This workload may be reduced by selecting candidates for study on the basis of probable gene function, SNP allele frequencies, haplotype block distribution and current studies of human: Chimpanzee gene orthology.

1. D. Bishop, Behavior Genetics (2001) 31, 4, 339-351

2. A. Klar, Genetics (2003) 165, 269-276

3. C. Francks et al., Am. J. Hum. Genet. (2002) 70, 800-805

4. C. Francks et al., Am. J. Hum. Genet. (2003) 72, 499-502

contact: c.phillips@mac.com 

O-26

Analysis of inter-specific mitochondrial DNA diversity for accurate species identification

Pereira F^1,2, Meirinhos J¹, Amorim A^1,2, Pereira L¹

¹ Instituto de Patologia e Imunologia Molecular da Universidade do Porto (IPATIMUP), Portugal; ² Faculdade de Ciências, Universidade do Porto, Portugal

Elucidation of several forensic casework studies relies on the precise identification of the species of origin for a variety of biological materials. With the advent of DNA based techniques, this correct identification has become of primary importance in different fields, such as in criminal investigations, food industry, protection of endangered species, etc. Nevertheless, the correct assignment of the biological samples sent to forensic laboratories has frequently proven to be a difficult task due to the high level of degradation and low quality DNA present in many samples as in the case of ancient materials (bone or teeth remains), stomach contents, hair, processed food (dairy products, roasted meat), etc. Several studies demonstrated that the information enclosed in the mitochondrial DNA (mtDNA) is the most useful and reliable tool for species identification, when compared with nuclear DNA based approaches, especially due to the large number of copies in each cell (raising the sensitivity of the analysis) and to the increasing number of sequences available in different databases. Commonly used mtDNA typing systems are based on the PCR amplification of a particular region of this molecule (usually the cytochrome b gene) followed by an RFLP or sequence analysis. However, attention should be paid to some points when using these techniques: (i) RFLP analysis are prone to false results due to undigested PCR fragments; (ii) the use of only one informative region may be not sensitive enough for the correct assignment; (iii) and the very difficult PCR amplification of larger fragments (> 300 bp) in old and/or degraded samples. In this work we attempt to develop a strategy to avoid some of these drawbacks and to produce more sensitive and reliable results for species identification. The first step was the construction of a large database, for different mtDNA regions, using all the available reference sequences for the class Mammalia (123 records; http://www.ncbi.nlm.nih.gov/genomes/ORGANELLES/40674.html). The alignment of these sequences allowed the correct identification of the diversity patterns found across the different mtDNA regions. We calculated these patterns of diversity splitting the aligned sequences in consecutive short windows of 100 bp overlapped by 50bp, using a home developed software. This characterisation will be useful for two main objectives: determination of the minimum fragment size suitable for typing highly degraded samples and the identification of regions for primer design. Conserved regions found in a wide range of species will be used for the design of primers that amplify segments containing species-specific information for species identification purpose. On the other hand, inter-specific variable regions are ideal for the design of primers for specific amplification in cases were mixed samples in very different proportions are suspected (a PCR with universal primers would lead to the identification of the most represented species only). Therefore, this information will be particularly useful in the development of a multiplex-PCR of short amplicons (≈100 – 130 bp) in different mtDNA regions for post-sequencing analysis, more informative and suitable for samples with degraded DNA.

Contact: fpereira@ipatimup.pt 

O-27

The Development of a DNA Analysis System for Pollen

Eliet J¹ and Harbison S²

¹Forensic Science Programme, Dep. Chemistry, Univ. Auckland, NZ; ²Forensic Biology, Inst. Environmental Science and Research Ltd, Auckland, New Zealand (SallyAnn.Harbison@esr.cri.nz)

Pollen is commonly identified as the yellow powdery substance that is found in flowers or released in large amounts from trees such as pines.  The use of pollen grains for forensic evidence was established in 1959 and has since been used successfully in many cases. The traditional method of pollen analysis is microscopy where the pollen grains are identified by the distinct patterns of the pollen wall. This is time consuming and requires a trained and experienced palynologist, of which there are few, and an extensive reference collection. This thesis aimed to develop a DNA analysis system for pollen grains extracted from soil that could be applied to forensic samples in casework. The focus on pollen grains recovered from soil was because a number of casework pollen samples are in the form of a soil sample, for example, mud from a shoe. DNA analysis techniques could provide advantages because it is quick and simple.  The DNA analysis investigated does not require in depth knowledge of the pollen types and reduces subjectivity associated with human judgment. The DNA analysis technique, terminal restriction fragment length polymorphism (tRFLP), was tested using the plant material from eight different species. Plant material was chosen to test the technique as plant material contains the same genes as pollen and the DNA extraction is relatively simple and effective. The DNA was extracted using a commercially available kit, the DNeasyÒ Plant Minikit (Qiagen). The tRFLP technique involved amplifying the extracted DNA using primers for the Adh1 gene, the forward primer labeled with FAM fluorescent dye. The amplification products were precipitated with ethanol prior to digestion with the Msp1 restriction enzyme. The restricted amplified product was analyzed by capillary electrophoresis on the ABI PrismÒ 3100 Genetic Analyzer (Applied Biosystems). As only the forward primer was labeled, the 5’ terminal restriction fragment was detected by the analyzer and this should be a different size for each different species. The results were interpreted using Genescan 3.7 software. The output values of peak area and fragment length were analyzed using a Euclidean distance measure to compare samples. The analysis showed that the tRFLP technique had good reproducibility as 93% of comparisons between replicates from the same species provided very strong support for them being the same species. The results also showed high discrimination between different species as the electropherogram profiles could be distinguished visually and the statistical analyses showed very high variation values for comparisons between samples from different species. Classification of an unknown as a particular species could be done correctly 97% of the time. A technique using glass bead maceration was found to be suitable for extracting amplifiable DNA from pollen grains. The pollen grain has a very tough wall made from a substance called sporopollenin. The force required to disrupt this wall, such as grinding in liquid nitrogen with mortar and pestle, is often too severe for the DNA to remain intact and results in damaged DNA, unable to be amplified. The glass bead maceration involved the addition of 1mm diameter glass beads to the sample with a sodium buffer and vortexing. This was sufficient to disrupt the pollen wall and release the DNA but did not result in the DNA being damaged. The same tRFLP technique was applied to the pollen DNA extract, as detailed above for the plant material.  Analysis of the DNA from five pollen species also indicated good reproducibility of the technique and discrimination between species. Pollen was seeded into soil samples to determine if soil had an effect on the extraction and amplification of DNA from pollen grains present in soil. Many chemicals used to remove pollen from soil for microscopic analysis are very harsh and remove everything inside the pollen grain including the DNA. Therefore, these methods are not suitable when DNA is to be extracted. The method of specific gravity separation with zinc bromide (ZnBr) was used to remove pollen grains from soil, as this was the least invasive method currently used. However, it was found to be unsuitable as the DNA extracted from the pollen grains that had been removed from the soil using ZnBr was degraded and non-amplifiable. An alternative method using sucrose was suggested, as it is less dangerous and should have no harmful effects on the DNA. In summary, the tRFLP technique was a reliable and reproducible technique that provides considerable discriminating power between samples. It will be suitable for application to forensic casework pollen samples after further work to improve the recovery of pollen from soil.

O-28

Characterizing Population Structure

Weir BS

Program in Statistical Genetics, Department of Statistics

North Carolina State University, Raleigh NC 27695-7566, USA

The population structure parameter theta, or Fst, is used in forensic match probability equations and also in results for parentage determination and remains identification. Although it describes the relationship among alleles within a population, it does so only with reference to alleles in different populations so that estimation requires data from more than one population. Standard methods for estimating theta provide an average over several populations and do not pay much attention to the sampling distributions of these estimates. A method for estimating population-specific values of theta will be described and illustrated with forensic STR data and for very dense SNP datasets.

Contact: weir@stat.ncsu.edu 

O-29

Autosomal Markers for Human Population Identification from Whole Genome SNP Analyses

Kayser M¹, Lao O^1,2, van Duijn JK^1,2, Kersbergen P^2,3, de Knijff P³

¹Department of Forensic Molecular Biology, Erasmus University Medical Centre Rotterdam, NL

²Department of Biology, Netherlands Forensic Institute, The Hague,

NL

³Department of Human and Clinical Genetics, Leiden University Medical Centre, Leiden, NL

Identifying the population of origin (=“ethnic origin”) of a perpetrator by DNA-analysis of a biological sample found at a crime scene would be highly useful for the police in order to concentrate their investigation on a specific group of individuals for finding an unknown suspect. For this purpose the use of sex-specific inherited Y-chromosomal and mitochondrial DNA markers is suitable due to their high degree of population affinity but the degree of confidence is limited by potential sex-biased genetic admixture. Therefore, autosomal markers are needed in addition to Y and mtDNA markers to identify the population / geographic region of genetic origin of an unknown individual with high degree of certainty.

We will present an approach for identifying informative autosomal markers for human population identification from whole genome SNP analysis. We have applied a whole genome scan including more than 10.000 single nucleotide polymorphisms (SNPs) in a set of globally dispersed human individuals and have used different statistical means to identify markers with maximal performance in population differentiation. Based on this dataset we have identified a small set of autosomal SNPs than can identify major human population groups. In order to test the capacity of those markers in other datasets we have typed them in a different set of human population samples including >50 regions from all over the world.

We want to emphasise that for those human populations showing a strong association with certain physical traits the genetic identification of the population of origin indirectly allows the prediction of externally visible characteristics (e.g. identification of African genetic origin predicts dark skin / hair /eye pigmentation). Thus, genetic markers for population identification, as presented here, will be the first attempt for predicting externally visible characteristics of an unknown individual by means of DNA-analysis, before a direct approach using markers that are functionally responsible for those phenotypic traits might be available in the future.

contact: m.kayser@erasmusmc.nl

O-30

A Compact Population Analysis Test Using 25 SNPs With Highly Diverse Allele Frequency Distributions

Phillips C¹, Sanchez J², Fontadevila M¹, Gómez-Tato A³, Alvarez-Dios³ J, Calaza M³, Casares de Cal, M³, Salas A¹, Ballard D⁴, Carracedo A¹ and The SNPforID Consortium⁵

¹Institute of Legal Medicine, University of Santiago de Compostela, Galicia, Spain; ²Department of Forensic Genetics, University of Copenhagen, Denmark; ³ Faculty of Mathematics, University of Santiago de Compostela; ⁴ Department of Haematology, Queen Mary’s School of Medicine, London, UK; ⁵www.snpforid.org

By selecting a total of 25 SNP loci that exhibit marked contrasts in allele frequency distributions (median highest frequency differential, d = 0.525 for 21/25 SNPs) in three major population groups: African, European and East Asian, we have developed a multiplex PCR assay and web based analysis tool that provides a predicted population of origin for a sample of unknown source. The genotyping assay was designed to use a single tube PCR and primer extension reaction and to be sensitive enough for routine forensic analysis. Appropriate markers were chosen from previously collected groups of population specific SNPs and non-binary SNPs (1, 2), from published ancestry informative marker sets, and from scrutiny of genes known to have been subject to diversifying selection in the recent evolutionary history of the population groups under study; e.g. FY in Africans and LCT in Europeans (3). The 25 SNPs comprising the final set were carefully selected to ensure as wide a distribution in autosomes as possible, maximising the potential for segregation of each marker. This is an important aspect of any population analysis test examining urban populations, since it can be expected that a large proportion of individuals from highly admixed populations or of immediate mixed descent (i.e. parental or grandparental), if undetected, would be incorrectly assigned to one of the contributing population groups.

SNP profiles generated from the genotyping assay can be submitted and analyzed with an open access web portal that uses a probability ratio approach based on the assumption of random variable independence for all markers. Three samples of 90 individuals each from Mozambique, Spain and Taiwan were used as training sets for the classification algorithm used. The error rate for a three population group classification was been estimated to be 2% from modelling (cross validation and bootstrapping) and below 1% from analysis of new profiles obtained from a different sample population in each group (60 Somali, Danish and Chinese samples).

(1) C. Phillips et al. (2004) Advances in Forensic Genetics 10, 233-235; (2) C. Phillips et al. (2004) Advances in Forensic Genetics 10, 27-29; (3) M. Jobling, M. Hurles, C.Tyler-Smith (2004) Human Evolutionary Genetics, Garland Science, New York

contact: c.phillips@mac.com 

O-31

A Bayes net solution that simulates the entire DNA process associated with analysis of short tandem repeat loci

Peter Gill¹,  James Curran²,  Keith Elliot¹

¹Forensic Science Service, Trident Court, 2960 Solihull Parkway, Birmingham, UK

²Department of Statistics, University of Waikato, New Zealand.

The use of expert systems to interpret short tandem repeat (STR) DNA profiles in forensic, medical and ancient DNA applications is becoming increasingly prevalent as high-throughput analytical systems generate large amounts of data that are time-consuming to process. With special reference to low copy number (LCN) applications we use a graphical model to simulate stochastic variation associated with the entire DNA process starting with extraction of sample, followed by the processing associated with preparation of a PCR reaction mix, and PCR itself. Each part of the process is modelled with input efficiency parameters ( ).  Then, the key output parameters that define the characteristics of a DNA profile are derived - namely heterozygous balance (Hb) and allele dropout p(D). The model can be used to estimate unknown efficiency parameters such as . ‘What-if’ scenarios can be used to improve and optimise the entire process - e.g. by increasing the aliquot forwarded to PCR the improvement expected to a given DNA profile can be reliably predicted.  We demonstrate that heterozygote balance and dropout are mainly a function of stochastic effect of pre-PCR molecular selection and can be predicted relative to the quantity of DNA analysed.  For mixture analysis, we show that the method is much more powerful than others suggested, since we simulate at the molecular level, without having to make assumptions based on a collection of output data (which may be unrepresentative).

We also show that whole genome amplification is unlikely to give any benefit over conventional PCR for LCN as there is no theoretical basis.

Contact: DNAPGill@compuserve.com

O-32

Maximisation of STR DNA typing success for touched objects

Prinz M, Schiffner L, Sebestyen J, Bajda E, Tamariz J, Shaler R, Baum H, Caragine T

Department of Forensic Biology, Office of Chief Medical Examiner, New York, NY

In order to produce database eligible DNA profiles from touched objects each individual step leading up to a DNA type was evaluated and optimized. The procedures were tested on fingerprints deposited on a variety of substrates, touched objects such as pens and credit cards, purified human embryonic kidney (HEK) cells with defined cell counts and diluted DNA from buccal swabs and other body fluids. For the initial swabbing several types of swabs and solutions were compared. DNA recovery was better for cotton fabric than for the conventional twisted thread cotton or Dacron swabs, while a  0.01% SDS solution performed better than water or other buffers. For the extraction, it was found that simple procedures with fewer steps were superior to commercial kits, such as DNA IQ^TM (Promega, Madison, WI) and QiaAmp (Qiagen, Valencia, CA), and other protocols with many manipulations.  The optimized protocol included a thirty-minute incubation with 0.01% SDS and proteinase K at 56^oC, followed by an incubation at 100^oC for 10 minutes.  Concentration of the extract and removal of the SDS was accomplished through centrifugation with a Microcon 100 (Millipore, Bedford, MA) column. The addition of 1ng Poly A RNA to the Microcon significantly improved DNA recovery. Samples were quantitated on a Rotorgene 3000 (Biotage) using an ALU repeat based real time DNA quantitation procedure as described by Nicklas and Buel (1). Based on work presented by Whitaker et al (2) samples were amplified in triplicate, with a minimum of 6.2 pg of DNA per amplification. Database compatible commercial megaplex kits were used for the amplification. For the Identifiler kit (Applied Biosystems, Foster City, CA) the annealing time was increased from 1 to 2 minutes and the cycle number was raised to 31 cycles. Initial experiments also involved Profiler Plus kits (Applied Biosystems, Foster City, CA) and the Poweplex 16 kit (Promega, Madison, WI). 6µL of amplified product were mixed with 15µLHiDi Formamide and 0.375µL LIZ size standard and analyzed on the 3100 Genetic Analzyer (all Applied Biosystems, Foster City, CA). Injection conditions were adjusted based on DNA input and three different conditions are being used routinely: 1kV 22 seconds, 3kV 20 seconds and 6 kV 30 seconds. Samples around 100pg gave the best results with 1kV 22 seconds, while 50 and 25pg samples were optimal at 3kV 20 seconds. The high injection conditions of 6kV, 30 seconds do result in broader peak shapes and but can be useful for the identification of low peaks. Peak intensities were maximised by not using variable binning and by setting the baseline window to 251. Data were analyzed empoying a minimum threshold of 75RFU. Alleles are only included in the interpretation if the allele is present in at least two of the three amplifications (2). The high injection conditions distort the expected peak intensities for low DNA amounts, therefore stochastic effects and allelic drop out events have to be newly characterized. Overall the increased cycle number and higher injection conditions allow reproducible DNA testing down to 20pg of DNA. For DNA dilutions, 25 pg routinely resulted in full profiles. For the touched objects, 78% of the 20pg to 100pg samples yielded database eligible profiles; the other samples were either mixtures or contained an insufficient number of allele calls. Here, the three amplification approach was crucial and yielded more complete profiles with more confidence in the allele calls. DNA amounts below 20pg did show partial profiles with correct allele calls that could have been compared in a specific case but were too incomplete for database entry.

Nicklas JA., and Buel E (2003) J. of Forensic Science  48: 282-291.

Whitaker JP, Cotton EA, and Gill P (2001) Forensic Sci.Int. 123: 215-223.

Contact: mprinz@nyc.rr.com

O-33

Multi-substrata analysis on Siberian mummies:

A different way for validation in ancient DNA studies?

Amory S ^1,2, Keyser-Tracqui C ^1,2, Crubézy E ², Ludes B ^1,2.

¹ Laboratoire d’Anthropologie Moléculaire, Institut de Médecine Légale, Strasbourg Cedex, France; ² Laboratoire d’Anthropobiologie, Université Paul Sabatier, CNRS, UMR 8555, Toulouse, France 31000.

Ancient DNA results are always submitted to caution due to the technical difficulties induced by the minute amounts, the degraded nature of the template and the high risk of contamination. A list of criteria of validation has been published as a guideline for ancient DNA researchers¹, including a dedicated and separated work area, controlled amplification, reproducibility of the results, etc… In addition to these criteria, the analysis of different substrates: bones, teeth and hairs of the same individual could be another way to ensure the reliability of the results.

This study presents the first results obtained on bones, molar teeth and hairs of two Siberian samples dated from the 18^th Century. Thus, the grave of Munur Urek, a burial site of an important clan chief and the multiple grave of the “Chamanic tree” site, gave us the opportunity to sample these different type of substrates. These two subjects excavated from frozen graves, were mummified. This exceptional state of preservation allowed us to test the amplification of autosomal and Y chromosome STRs and the sequencing of the HVI region on the three types of substrates. All experiments were done in a dedicated laboratory and negative controls were run for each step. The persons in contact with the samples were typed for the same markers in order to determine exogenous contamination.

This method permitted the identification of artefacts on STRs profiles, common when working with Low Copy Number amounts of DNA. Indeed, the comparison of the profiles obtained for bones and teeth highlights allelic dropouts and spurious alleles for the bone samples.

The possibility to compare results from different substrates, in spite of the limited numbers of possible cases, represent another, and interesting, criterion to confirm the authenticity of ancient DNA results.

¹Cooper A, Poinar HN. Ancient DNA: Do It Right or Not at All. 2000. Science, Vol 289, Issue 5482, 1139.

Contact: Sylvain.Amory@iml-ulp.u-strasbg.fr