Incest indices from microsatellite genotypes of mother-child pairs

Suspected incestuous paternity is encountered infrequently and investigation may be complicated by absence of the suspected father. Incest indices (IIs) can be calculated from microsatellite (STR) types of only a mother and child, but could be misleading. Therefore, the method was evaluated.

Combined incest indices (CIIs) of 50 randomly mated (RM) mothers and their children were compared with those of 50 simulated incestuous (SI) mothers and their children. Each CII was calculated from 18 individual locus IIs. Combined indices were categorized as "diagnostic" (<0.010 and >100 for RM and SI cases, respectively), "indicative" (CII was directionally correct but not erroneously diagnostic), and "misleading" (>1.0 in RM and <0.01 in SI). The relative importance was determined of each of the three variables contributing to the CII.

In 41 cases (41%), CIIs attained diagnostic values. Fifty-two CIIs were indicative. CIIs were misleading in 3 RM cases and 4 SI cases. The number of mother-child (M-C) STR genotype similarities was the most important determinant of CIIs. Infrequent alleles in M-C similarities were important in raising CIIs in SI. The kind of M-C genotype similarity was the least important variable.

Study of 18 STR loci produces diagnostic CIIs in only two of five suspected incest cases. Study of approximately 33 independent STRs would assure that greater than 97.5 percent of cases will have diagnostic CIIs if incest occurred. Study of loci that are more informative than typical STRs would be advantageous.