Maternal typing and test sufficiency in parentage analyses.

BACKGROUND: The contribution of maternal typing to paternity analysis was evaluated to determine how many additional loci to study in one-parent cases. STUDY DESIGN AND METHODS: Four groups underwent paternity analyses with an eight-locus test battery. Files of 25 case trios were retrieved, in which alleged fathers had achieved paternity indices of greater than 100 ("included trios"). Maternal types were omitted and the cases were reanalyzed ("included duos"). Mother-child pairs of the cases were then coupled with unrelated men ("excluded trios"), and the cases were analyzed. Maternal types were omitted from the excluded trios and cases were reanalyzed ("excluded duos"). RESULTS: Paternity indices of men in included duos were markedly reduced when compared to included trios; odds were sufficiently low in 9 of 25 men that paternity remained in doubt. After omission of maternal phenotypes, excluded duos exposed 33 percent fewer genetic inconsistencies than excluded trios; 5 of 25 men in excluded duos demonstrated less than two genetic inconsistencies and 1 man had none. The specific probabilities of paternity exclusion in motherless cases averaged 61 percent per locus of those in case trios. One random man in 52 duos was not excluded by the eight tests versus 1 in 417 trios. CONCLUSIONS: Omission of maternal typing from eight common microsatellite paternity tests reduced conclusive evidence for or against paternity by 30 to 40 percent. False inclusion of random men is an important failing of tests in motherless cases. Cases involving one parent and child (e.g., in immigration) would require examination of an additional five similar loci to compensate for absent maternal data. A change in standards is suggested.

The specific power of parentage exclusion in a child's blood relatives.

BACKGROUND: The impersonation of parent and child by two other blood relatives is an important problem in parentage analysis involving potential immigrants. STUDY DESIGN AND METHODS: A statistic (AR) is proposed, based on the specific power of exclusion of paternity, which describes the ability of a child's test results to demonstrate evidence of nonparentage under the hypothesis that an ostensible parent is actually an older sibling. A case illustrates the value of A(R): a woman and her two alleged children were typed at 3 variable number of tandem repeat (VNTR) loci after 18 short tandem repeat (STR) loci initially showed strong evidence of the woman's maternity of one child and her exclusion from parentage of the second. AR and 1 - AR were calculated from the STR types of the first child. RESULTS: The woman was excluded from maternity of both children with the additional VNTR tests. Given the 18 STR test findings of the first child, the probability was 12 percent that there would be no inconsistencies with parentage in a sibling pretending to be a parent. CONCLUSION: The value 1 - AR, siblings not excluded from parentage, explains how a seemingly large number of examined loci can fail to reveal even one genetic inconsistency if two siblings have posed as parent and child. Approximately 25 STR loci appear necessary to achieve 95 percent confidence of detecting at least one genetic inconsistency indicative of nonparentage.