Discovering Kinship through Small Subsets.

In kinship inference, we identify genealogical relationships among organisms. One such problem is sibgroup reconstruction: given a population of same-generation individuals, partition it into sibgroups resulting from mating events. Minimizing the number of matings is NP-hard to approximate, yet a simple heuristic, based on identifying population triplets that can be from the same sibgroup, performs comparably to better than integer programming algorithms in a fraction of the running time. With high probability if we study many loci in the genome, and large populations, our polynomial-time heuristic finds the true sibgroups, assuming a standard probabilistic inheritance model. [ABSTRACT FROM AUTHOR]