Evaluating factors that affect full- and half-sibling inferences via genetic pedigree reconstruction without parental genotypes

Abstract

Close-kin mark-recapture (CKMR) is a challenging but exciting new approach to study dispersal because sampled and genotyped individuals can be used to “tag” full- and half-sibling family members without using parental genotypes via sibship reconstruction. CKMR increases sampling efficiency and, as a result, has enabled abundance estimation of migratory species and improved understanding of dispersal for species that are difficult to sample using traditional methods. Accordingly, there is interest in using CKMR to study dispersal ecology for the invasive sea lamprey in the Great Lakes, yet much is unknown about this new approach. We investigated the performance of a Rapture panel for sea lamprey reconstruction pedigrees without parental genotypes using two programs, Colony and Sequoia. Initially, datasets were created in the simulation module of Colony, where the number of parents (250, 500, 1000), number of offspring (250 or 500), and the number of loci (400, 600, 800) varied totaling 50 datasets for each of the 18 parameter combinations. Colony and Sequoia were then used to reconstruct the pedigrees without parental genotypes. Performance metrics were compared across parameter space and between sibship reconstruction programs. Colony and Sequoia inferred full-sibling and unrelated individuals with low assignment error. Sequoia was much faster at reconstructing pedigrees (<5 min compared to ∼ 1 week); however, the program nearly always (>0.99) misassigned known half-siblings as unrelated, which is an issue because sea lamprey likely produce many half-siblings. Accordingly, Colony should be used to reconstruct sea lamprey pedigrees when only offspring genotypes are provided.