Determination of sibship in any two persons

Abstract

Parentage testing laboratories may be asked to provide genetic evidence that two persons are or are not related, when no other relatives are available for study. Simple methods using autosomal, codominant, unlinked genetic systems can determine if two people are blood relatives (e.g., siblings). The odds ratios (full sibship index) of true sibling pairs were determined from two-child paternity cases and compared with regionally and racially matched control pairs of unrelated children. The sharing of two, one, or no alleles was observed in pairs of children at three independent, polymorphic VNTR (variable number of tandem repeat) sequences loci. The sibship index was calculated as (the chance that an observation would occur if two children were siblings) divided by (the chance that it would occur if the two were unrelated). Sibship indices and the frequencies of shared alleles were determined for 20 sibling pairs and 20 control pairs. Sibship index values were less than 1 in all 20 pairs of unrelated children. Sibship index values were greater than 100 in nine pairs of siblings (45%), between 10 and 100 in five pairs (25%), between 1 and 10 in four pairs (20%), and less than 1 in two pairs (10%). Siblings shared two alleles in 17 of 60 observations (28.3%); controls shared two alleles in 0 of 60 observations (0%). The sharing of one allele and the sharing of no alleles at a polymorphic locus of high heterozygosity provide limited information for and against sibship, respectively. The sharing of two alleles produces strong evidence favoring sibship. In a given case, the study of more than three polymorphic loci of high heterozygosity may be needed to develop the evidence that two people are siblings. The general logic and methods used for siblings apply to kinship analyses of other two-person pedigrees.