Two-locus and three-locus gene identity by descent in pedigrees.

Although there have been several mathematical formulations of multilocus segregation, multilocus gene identity by descent in pedigrees has been little considered. Here we present a computationally feasible algorithm for the computation of two-locus kinship for individuals between whom there may be multiple complex relationships, and use it to investigate patterns of two-locus gene identity by descent for some standard relationships. We also present an explicit formula, which is used to discuss the determinants of two-locus identity and the relationship to three-locus identity by descent. With the current increasing density of information on individual genomes available from DNA polymorphisms, gene identity at linked loci has practical importance. Procedures for the estimation of relationships between individuals on the basis of genetic data will have increased flexibility to discriminate wider classes of genealogical relationship where information on multiple linked loci can be employed. Gene identity by descent at linked loci is also a key aspect of mapping rare recessive diseases from data on inbred individuals.