Terry Melton

Genetic evidence for the proto-Austronesian homeland in Asia: mtDNA and nuclear DNA variation in Taiwanese aboriginal tribes.

Previous studies of mtDNA variation in indigenous Taiwanese populations have suggested that they held an ancestral position in the spread of mtDNAs throughout Southeast Asia and Oceania (Melton et al. 1995; Sykes et al. 1995), but the question of an absolute proto-Austronesian homeland remains. To search for Asian roots for indigenous Taiwanese populations, 28 mtDNAs representative of variation in four tribal groups (Ami, Atayal, Bunun, and Paiwan) were sequenced and were compared with each other and with mtDNAs from 25 other populations from Asia and Oceania. In addition, eight polymorphic Alu insertion loci were analyzed, to determine if the pattern of mtDNA variation is concordant with nuclear DNA variation. Tribal groups shared considerable mtDNA sequence identity (P>.90), where gene flow is believed to have been low, arguing for a common source or sources for the tribes. mtDNAs with a 9-bp deletion have considerable mainland-Asian diversity and have spread to Southeast Asia and Oceania through a Taiwanese bottleneck. Only four Taiwanese mtDNA haplotypes without the 9-bp deletion were shared with any other populations, but these shared types were widely dispersed geographically throughout mainland Asia. Phylogenetic and principal-component analyses of Alu loci were concordant with conclusions from the mtDNA analyses; overall, the results suggest that the Taiwanese have temporally deep roots, probably in central or south China, and have been isolated from other Asian populations in recent history.

Extent of heterogeneity in mitochondrial DNA of European populations

Variation in the mitochondrial DNA (mtDNA) control region as detected by sequence-specific oligonucleotide (SSO) probes is described for 595 individuals from six European or European-derived populations. Estimates of diversity for mtDNA types exceed 0.91 in all populations, while 50% of the 158 types which were observed occur only once. Of 68 shared types, most occur rarely (< 3% of the total population); only one type occurs at a frequency greater than 10%, and it is present at comparable frequencies in all six populations (18-29%). An analysis of molecular variance (AMOVA) incorporating genetic distances between types shows that 100% of the variation present in the total sample is attributable to within-population diversity, while there are essentially no between-population differences. Another AMOVA was performed for the first hypervariable region SSO sites only, which included this sample plus an additional 537 SSO types from mine more European populations that were inferred from published mtDNA control region sequence data. Similar results were obtained, with over 99% of the variation overall attributable to within-population differences, and less than 1% of the variation attributable to between-population differences. The Saami were the most different from other populations, which had been observed in an earlier study of nucleotide sequence data. Overall, there is no statistically significant heterogeneity for European populations (p > 0.001), and these groups are virtually indistinguishable with respect to mtDNA SSO types. These results demonstrate the utility of mtDNA typing for forensic investigations.

Forensic Mitochondrial DNA Analysis of 116 Casework Skeletal Samples

ABSTRACT: Between February 1999 and May 2005, 116 DNA extractions were completed on skeletal remains from routine casework. Overall, at least a partial mitochondrial DNA (mtDNA) profile was obtained on 83.6% of samples. Skeletal remains fell into two general categories: (1) samples for body identifications submitted by law enforcement and (2) samples submitted to answer historical or family identity questions. Body identification cases were more likely to yield full mtDNA profiles, whereas historical cases were more likely to result in partial profiles. Overall, the ability to obtain a full or partial profile primarily reflects the difference in the average age and condition of the samples in these two categories and thus, difference in the quantity and quality of the DNA. Cremated remains were uniformly unsuccessful, whereas infant/fetal remains were uniformly successful. Heteroplasmy in skeletal remains was observed at a rate similar to that in hair (∼10%). For body identification cases, skeletal remains had the same mtDNA profile as the accompanying reference sample in 50% of cases.

A cautionary note on switching mitochondrial DNA reference sequences in forensic genetics

The first human mitochondrial DNA (mtDNA) sequence was produced in 1981 from an individual of European descent [1]. Since then, this sequence has been known as the Cambridge Reference Sequence (CRS) with a total length of 16,569 base pairs. As is common practice in other fields of genome research, this first mitochondrial genome (mtGenome) served as reference for the scientific community, relative to which other mtDNA haplotypes were reported. Eighteen years later the CRS was re-sequenced and corrected at 10 positions (3423T, 4985A, 9559C, 11335C, 13702C, 14199T, 14272C, 14365C, 14368C, and 14766C) to form the revised Cambridge Reference Sequence (rCRS) [2]. The new analysis also revealed that this mtGenome consists of only 16,568 nucleotides, as a base at position 3107 was mistakenly reported in the CRS. Instead of redefining all nucleotide positions downstream of 3107, this position is indicated in the rCRS as a deletion (unfortunately often indicated as ‘‘N’’, which is reserved for any base in the IUPcode [3]). Thus, the numbering system employed for the CRS and the body of established data can continuously be used with the rCRS. More than 15,500 mtGenomes and well over 150,000 (partial) control region sequences (including databases) have been published to date (http://www.phylotree.org/, Phylotree Build 14 [4]), in which the CRS and the rCRS have been cited 5603 and 968 times, respectively (http://apps.isiknowledge.com/; queried on May 2012). In a recent study, the switch to a new reference sequence, the so-called Reconstructed Sapiens Reference Sequence (RSRS), has been proposed [5]. This ancestral reconstructed sequence represents the deepest root in the known human mtDNA phylogeny at the base of the split of haplogroups L0 and L10203040506 after combining sequence information from all available mtGenomes from Homo neanderthalensis and novel human mtGenomes. The authors believe that the switch would solve misunderstandings and problems associated with the existing nomenclature relative to the rCRS, which belongs to the recently coalescing European haplogroup H2a2a1. In the following we briefly review developments in forensic mitochondrial genetics and discuss possible implications of the proposed switch. Mitochondrial DNA is highly abundant in cells compared to nuclear DNA (nDNA) with increased typing success rates for analysis of highly degraded samples and also hair shafts that often do not harbor detectable amounts of nDNA. To assess the significance of a match between two mtDNA haplotypes (e.g. from the crime scene and from a suspect), mtDNA databases have been developed. Earlier, the difference-coded haplotypes (with respect to the rCRS) were directly compared to the haplotypes in the mtDNA databases with the risk that multiple different alignments of the same sequence led to biased results [6]. The

Diversity and Heterogeneity in Mitochondrial DNA of North American Populations

Variation in the mitochondrial DNA (mtDNA) control region as detected by sequence-specific oligonucleotide (SSO) probes is described for 2282 individuals from African-American, European-American, and Hispanic subpopulations from five broadly defined regions of North America (Northeast, Southeast, Central, Northwest, Southwest). Population diversity estimates were uniformly high for all subpopulations and for each major ethnic group. Only the Pennsylvania Hispanic group was remarkable with respect to its mitochondrial DNA types, having both six low frequency population specific types (ranging from 1.2-8.6%) and three high frequency shared types (10-20% each). There was no statistically significant subpopulation heterogeneity present within any of the three major groups at either the subpopulation level or the regional level (p > 0.01). However, statistically significant heterogeneity was measured when comparing the three major groups to each other, with the variance component attributable to this large division accounting for 18.60% of the total variance (p < 0.001). Overall mtDNA is a satisfactory forensic typing locus within broadly defined African-American, European-American, and Hispanic groups from North America, based on the high diversity estimates and absence of heterogeneity, as characterized by SSO typing.

EXTENT OF HETEROGENEITY IN MITOCHONDRIAL DNA OF SUB-SAHARAN AFRICAN POPULATIONS

Variation in the mitochondrial DNA (mtDNA) control region as detected by sequence-specific oligonucleotide (SSO) probes is described for 381 individuals from nine sub-Saharan African populations. Population diversity estimates for SSO types ranged from 0.23 to 0.97, while 102 SSO types were detected, none of these types was shared by more than four populations. Eighteen types occurred in > or = 10% of individuals in some populations; of these, 11 were population-specific. One type occurred in 15% of the total sample, but was shared among only three populations. African SSO types were characterized by high frequencies of blank variants, indicating that there was additional variation present at the nucleotide sequence level in regions where SSO probes hybridize. Analyses of molecular variance (AMOVA) incorporating genetic distances between SSO types showed that 30% of the total variation was due to differences among populations, indicating that there is statistically significant heterogeneity (p < 0.001). An AMOVA on mtDNA control region nucleotide sequence data from 12 populations showed that including all additional variation present at the sequence level increased the variance due to population subdivision to 34% (p < 0.001). Overall, when considering both the low diversity within some populations and high heterogeneity among populations, SSO typing of mtDNA may not be a desirable forensic DNA typing method for continental African populations. Further mtDNA sampling of African-derived populations of North America should be carried out to determine how much of the continental African mtDNA variation is of forensic significance. However, the existence of extensive mtDNA control region nucleotide sequence variation in African populations means that control region sequencing is still appropriate in forensic cases requiring mtDNA analysis.

Extent of heterogeneity in mitochondrial DNA of ethnic Asian populations.

Variation in the mitochondrial DNA (mtDNA) control region as detected by sequence-specific oligonucleotide (SSO) probes is described for 993 individuals in 11 ethnic Asian populations. Estimates of diversity for mtDNA types exceed 0.94 in all populations, while 53% of the 255 types that were observed occur only once. Of 96 shared types, four occur at frequencies of greater than 10% but less than 17% in any one population. There is statistically significant heterogeneity among these 11 populations, however, an analysis of variance incorporating genetic distances between types shows that at least 95% of the variation present in the total sample is attributable to within-population diversity, while only 5% is due to between-population differences. Overall, heterogeneity with respect to mtDNA SSO types is grossly correlated with geographic distance between populations; the most extreme heterogeneity was observed between populations from East Asia and populations from West Asia. With respect to population genetics, the control region of mtDNA exhibits satisfactory qualities as a DNA typing locus.

Mitochondrial DNA analysis of 114 hairs measuring less than 1 cm from a 19-year-old homicide

BackgroundMitochondrial DNA analysis is typically applied to degraded skeletal remains and telogen or rootless hairs. Data on the application of the method to very small hairs less than 0.5 cm from an age-matched and -challenged sample set are lacking.MethodsOne hundred fourteen hairs sized less than 1 cm from a 1993 case were analyzed for mitochondrial DNA according to laboratory standard operating procedures. For some hairs, a screening approach was applied, which permitted some samples, such as victim hairs on victim clothing, to be eliminated from the process quickly. Degraded samples were amplified with “mini-primers,” and 12S species testing was applied when non-human hairs were encountered.ResultsPartial to full control region human mitochondrial DNA profiles or species identifications (non-human species) were obtained from 93% of hairs under 1 cm, 92% of hairs under 5 mm, and 90% of hairs under 3.5 mm. Nineteen of 21 hairs 2 mm or less gave full or partial profiles. Among 128 hairs of all sizes tested in the case, 9 gave no results, 3 were canine in origin, and 73 did not exclude six known individuals tested in the case. Twenty-two hairs had nine additional profiles that were observed two or more times each. Twenty-one hairs showed singleton types not matching each other or any individual.ConclusionsCrime scene hairs that are both aged and small are often judged to be unsuitable for either hair microscopy or DNA analysis. This study of age-matched challenged small hairs indicates that even the smallest probative crime scene hairs are suitable for mitochondrial DNA analysis and can provide useful data.

Commentary on: Divne A-M, Nilsson M, Calloway C, Reynolds R, Erlich H, Allen M. Forensic Casework Analysis Using the HVI/HVII mtDNA Linear Array Assay. J Forensic Sci 2005;50:548–54.

Sir: We read with interest Divne et al.’s ‘‘Forensic casework analysis using the HVI/HVII mitochondrial DNA (mtDNA) linear array assay’’ (1). As caseworking mtDNA forensic examiners with hundreds of mtDNA cases behind us, including experience with hundreds of biological specimens of all kinds (2–4), we urge caution in the use of any mtDNA screening method like the linear array assay that develops only a partial profile on evidentiary material. One of us (T. M.) also has had extensive experience with the precursor technology of linear arrays, SSO typing [see, for example, (5)] and has a good idea of its limitations, especially the potential for a high frequency of ‘‘null’’ or ‘‘blank’’ results caused by polymorphisms that block hybridization, a problem which is barely mentioned by Divne et al. Our forensic concerns are in these general areas:

Commentary on: Foran DR, Gehring ME, Stallworth SE. The recovery and analysis of mitochondrial DNA from exploded pipe bombs. J Forensic Sci 54;1:90-4.

Sir, Foran et al. (JFS 54:90–94) describe mitochondrial DNA (mtDNA) analysis of handled pipe bomb components. The practices used in their experiments are inappropriate in forensic casework. Ninety-four of 114 PCR amplifications of the swab samples in this experiment yielded insufficient or no amplification product for DNA sequence analysis after 38 cycles of mtDNA amplification. These samples were subsequently processed further: a 1 lL subsample of first-round PCR product, whether it was visible on a yield gel or not, was placed into a second nested amplification of 24 cycles, giving a total of 62 cycles of amplification. With this approach, every swab produced an amplification product. Due to tight controls applied to handling of materials prior to sampling, and because the investigators knew the profiles of all contributors, every amplification product should have been assigned to a donor. However, the investigators recovered 10 profiles (28% of the samples) that could not be assigned. Therein lies the problem of nested PCR. The use of nested PCR can yield amplification product from contaminant molecules rather than the target DNA, particularly when the amount of target DNA is minimal. If a contaminant does become detectable through nested PCR, it is impossible to discriminate between the contaminant, a handler, or someone who may have handled the materials some time well before bomb preparation. In addition, PCR artifacts may result due to stochastic effects that occur when amplifying low levels of DNA. Previous descriptions (1,2) of the use of nested PCR in the examination of mtDNA have also demonstrated increased risk of contamination and elevated background noise. In fact, Gryzbowski et al. (2) state that ‘‘...the nonreproducibility of the results...suggest that some of those mutations might be artifacts resulting from specific conditions of nested PCR...,’’ ‘‘...specific conditions of nested PCR favor the occurrence of PCR replication errors...,’’ and ‘‘...the exclusion of nested PCR from the techniques employed in forensic casework would be a more conservative approach....’’ Brandstatter and Parson (3) also found that higher quality electropherograms were obtained from direct sequencing when compared with nested PCR products. The ancient DNA and forensic DNA communities have been aware of these potential problems for nearly two decades, and the use of more than 40 cycles of PCR amplification in mtDNA analysis and 34 cycles of PCR amplification in short tandem repeat (STR) analysis has been rejected by the broad forensic community. In addition, while Foran used ‘‘control samples’’ that were not handled by any subjects participating in the experiments, there is no indication that simple reagent blank extraction negatives were run through the comparable 62 cycles of PCR, or if so, what they yielded. The authors do not report the results of the PCR negatives described in the experimental methods. While the control bomb samples gave ‘‘nonsense’’ data, as opposed to negative results, there is no description of what these sequences consisted of, or that they were investigated for a possible origin. The negligible difference in the nested fragment sizes (256 bp vs. 283 bp) seems unlikely to explain the difference in obtaining results for HV1 and HV2. However, first round amplification products were significantly different in size: 403 bp (HV2) versus 333 bp and 266 bp (HV1). It is unlikely that forensic evidence samples could routinely and successfully be amplified in a 403 bp amplicon. In fact, standard forensic practices such as the use of smaller amplicon sizes and even use of mini-primers might have been suitable for this evidence without nested PCR. While mixtures, with their overwhelming interpretational challenges, might have been obtained on some samples, the use of smaller amplicons might have yielded single profiles on some samples that were not successful at 38 cycles with the larger amplicons. However, given the sensitivity of mtDNA analysis and our experience with its use on touched objects, we would predict a high incidence of detecting mixtures on touched casework items, which have likely been exposed to and handled by numerous individuals. Furthermore, it is well known in both the forensic and global science community that the peaks generated by Dye-Terminator sequencing chemistry are not quantitative. With current technology, the practice of attempting to assign peaks to an individual when a mixture is present in a sequencing trace is limited, especially in a circumstance such as this one where experiments have been performed on DNA amplified under conditions sure to induce stochastic effects. Additionally, only a single-strand of sequence data containing high levels of background noise was presented by the authors, illustrating the problems of interpreting such data. Finally, it should be noted that the authors soaked or cleaned the bomb components in a 10% bleach solution prior to perform experiments. While a controlled study should first be performed with the cleanest samples possible, this study shows that the cleaning steps taken in these experiments still gave erroneous results. Given that bomb components obtained from a true crime scene would have a far greater potential for contamination than that seen in these experiments, the problems seen here would be exacerbated under real world conditions. The authors also reported an ‘‘individualizing success rate, given the closed population,’’ as 50%. An analysis performed at this sensitivity cannot assume a closed system and the reported 28% of profiles that were unassignable in these experiments demonstrate that this type of testing cannot individualize the source of evidentiary items. In fact, the great strength of mtDNA analysis is its ability to exclude individuals as potential sources of evidence. The investigators state that DNA quality is the relevant factor for obtaining results. As the fragment sizes here are similar to those of small STR amplicons, we maintain that copy number is one critical factor in obtaining successful results. When copy number is a critical feature, the ability to amplify a single contaminant molecule with 62 cycles of PCR to a detectable level will result in meaningless outcomes. Another critical factor in obtaining DNA profiles from any type of evidence is the absence of inhibitors. The probable presence of explosive residue and its impact on the PCR cannot be overlooked as an important parameter in the analysis of postblast DNA. The assertion that this study ‘‘holds more promise than any technique that has preceded it’’ for identification of IED assemblers should be viewed with extreme skepticism. J Forensic Sci, July 2009, Vol. 54, No. 4 doi: 10.1111/j.1556-4029.2009.01082.x Available online at: www.blackwell-synergy.com