How high should paternity index be for reliable identification of war victims by DNA typing?

Aim. To analyze statistically and logically the significance of genetic matches between skeletal remains and relatives of missing persons in the process of identification of war victims by DNA typing. Methods. DNA was isolated from bone and blood samples and short tandem repeat (STR) loci were typed by using AmpFLSTR Profiler, Profiler Plus, and Identifiler kits. Novel mini-haplotype analysis that compares matches in all three-locus combinations of alleles was developed and used in the analysis of inbreeding in the group of 295 unrelated individuals. Results. While comparing 98 skeletal remains exhumed in the process of identification of war victims in Croatia with over 3,000 genotypes of relatives of missing persons, we revealed 20 cases of 14-locus matches and 4 cases of 15-locus matches between unrelated people. We hypothesized that this unexpectedly high number of false matches might be a consequence of local inbreeding and supported this hypothesis with very. low correlation between the probability of a genotype and the number of matching genotypes in the database (R-2 = 0.36). Further support for the hypothesis was obtained by the analysis of mini-haplotypes, which revealed up to 90% overrepresentation of some mini-haplotypes. Conclusions. STR DNA typing is the "golden standard" of human identification, but evidential value of a genetic match can be easily misinterpreted. Therefore, careful use of statistical methods is essential for the proper evaluation of laboratory results. Whenever possible, multiple relatives should be analyzed and other evidence based on the information about time, place, and other conditions of disappearance, as well as anthropological and other "classical" forensic data should always be put together and compared before any final decision about the identity is made.