The application of DNA profiling to the identification of victims in the Gulf war (1990-1991)

Abstract

This project was designed and developed in response to the need to improve the methodology employed in the DNA profiling of the Kuwaiti victims of the First Gulf War (1990-1991). The main challenges have involved developing the
methodology in an attempt to increase the DNA recovery from the skeletal remains and also assess the preservation of DNA in the remains. In addition, work was undertaken to assess two commercial STR amplification kits, the Identifiler® and MiniFilerTM, to establish allele frequency databases for use in Kuwait and to assess the concordance of the two kits.
In order to assess the methodology for DNA extraction and prediction of DNA preservation two sources of materials were used: simulated casework samples and actual casework samples. To obtain simulated casework samples, bones
from sheep and teeth from human were buried at three different sites within Kuwait. These were sampled over a period of 60 weeks. In addition, samples from the femur and humerus of 25 individuals who were killed during the Gulf War, but had not yet been identified, were taken for analysis. These were exhumed from five gravesites, three in Iraq and two in Kuwait. Previous attempts to generate DNA profiles from the samples had failed.
Different extractions protocols and purification methods were assessed including: a phenol:chloroform-based extraction; the GENECLEAN® Kit (Obiogene); QlAquick Gel Extraction kit (Qiagen); the QIAamp DNA Blood Maxi kit (Qiagen), using a protocol based on Davoren et al (2007); and a modified silica-based extraction using the DNeasy® Blood and Tissue Kit (Qiagen). PCR amplification of the extracts and the real-time quantification results showed that the modification of a silica-based method, using the Qiagen DNeasy® kit, was successful in removing inhibitors that were present in the extracts and obtained with all the other extraction methods. This allowed the successful profiling of 19 out of the 25 samples that had previously failed.
In an attempt to further improve the DNA extraction efficiency, the effect of Nphenacylthiazolium bromide (PTB) was assessed. PTB has been reported previously to improve DNA extraction from ancient DNA samples (Paabo, 1989;
Poinar et al., 1998) by releasing DNA that has become cross-linked with proteins. In this study, the effect of PTB, while statistically significant when used with samples from some sites, was minimal.
The power of different methods to allow an effective system of triage (sorting of samples based on the likelihood of successful analysis) was examined. Three parameters were assessed: gross morphology, histology, and chemical status
of the bones were compared with the amount and quality of DNA that was recovered from different samples. The simulated casework samples displayed only minor changes in gross morphology and histology over the period of the study, whereas the casework samples displayed varying degrees of change.
The samples from Iraqi sites generally displayed good morphological and histological preservation. In contrast, the samples from the two sites within Kuwait displayed an almost complete lack of histological features and changes (pitting/cracks) to the surface. The morphological and histological preservation correlated closely with the success rate when extracting DNA from casework samples that were buried in Iraq and Kuwait. Nitrogen content in all samples was very similar and the results showed that it was not a useful indicator of preservation.

The MiniFilertM (Applied Biosystems) is designed for the analysis of degraded
DNA. Before applying this to casework, it is important to carry out a
concordance study in order to ensure the results with the MiniFilerTM are
comparable to the Identifiler® (Applied Biosystems) DNA profiles. The reference
database with relatives' DNA profiles are all generated using Identifiler®. To
assess the concordance, the MiniFilerTM profiles from 200 unrelated Kuwaiti
samples were compared to Identifiler® profiles. Concordance was observed for
99.875% of the compared loci (1598 of 1600). The two discordant profiles
displayed allelic dropout: one at the Dl 35317 locus due to non-amplification of
allele 10 in the MiniFilerTM profile, and one at the D18S51 locus due to nonamplification
of allele 18 in the Identifiler® profile.
Finally, since the population of Kuwait is heterogeneous, with a strong tribal
system, the possibility of subpopulation effect within the Kuwaiti population was
examined. Allele frequencies for the 15 STR loci included in Identifiler® kit were
ascertained in a sample population of 502 unrelated Kuwaiti individuals. The
results were compared with 6 different populations. The Kuwaiti population was
very similar to neighboring Iraqi and Saudi populations. These data are now
used in casework undertaken in Kuwait, to calculate the statistical significance
of matching DNA profiles (the results of the reference database work are
included in Appendix 1 rather than in the main body of the thesis).