The identification of war victims by reverse paternity is associated with significant risks of false inclusion.

Since February 2001 the process of DNA identification of war victims in Croatia relies on the database of over 3,000 9-locus (D3S1358, vWA, FGA, TH01, TPOX, CSF1PO, D5S818, D13S317 and D7S820) STR genotypes of relatives of missing persons. Instead of a targeted approach to DNA typing, the genotype of each skeletal remains analysed is compared to all genotypes in the database to identify potential parents and children. Although this approach has significantly increased the pace of identification by DNA typing, non-targeted matching in a database containing several thousand genotypes considerably decreases the significance of inclusion, especially when identification is based on reverse paternity analysis. To support this statistical prediction we present 3 cases of 10 STR loci matches and 1 case of 11 STR loci matches between a child, child's mother and skeletal remains that did not originate from a father of that child.