Use of sibling pairs to determine the familial searching efficiency of forensic databases

Abstract Pairs of individuals tested at the 13 CODIS core STR loci to determine sibship were used as a source of familial data that was seeded into a larger data set of 12,000 plus DNA profiles simulating a CODIS-like offender database. To determine whether known sibs could be found in the larger database two methods were used: degree of allele sharing and a kinship matching approach. The allele sharing method detected 62 of 109 of the known sib pairs (57%) while kinship matching detected 90 of the sib pairs (83%). Although kinship matching was the more efficient method of the two, the number of false positives generated prior to finding a true match was inversely related to the likelihood of sibship suggesting that many true siblings would not be easily found in a large forensic database via familial searching techniques.