Assignment of Individuals to Populations: Bayesian Methods and Multi-Locus Genotypes

Abstract

Multi-locus genotype data have become a powerful management tool to identify degree of population structure, assign individuals to populations, identify hybridization, and assess rates of migration between populations. We use an example from the tuberculosis (TB)-infected white-tailed deer (Odocoileus virginianus) population in northeast Michigan (Deer Management Unit [DMU] 452) to illustrate the utility of multi-locus genotype data combined with assignment testing approaches in a wildlife management setting. We used microsatellite loci to establish that an unknown TB-positive deer tissue sample originated from a deer outside DMU452. Subsequently, we used multi-locus genotypic data and a likelihood-based assignment test to place TB-positive deer harvested outside of DMU452 into their populations of origin. We used genetic algorithms to identify a maximally informative subset of loci. We also determined estimates of statistical confidence in individual assignment decisions. Data indicate that the unknown TB-positive deer tissue sample was from a deer harvested outside DMU452. Two TB-positive deer killed outside of DMU452 were assigned to the local populations where they were harvested. The presence of TB-positive deer outside DMU452 suggests that TB is present in multiple areas of Michigan, dictating increased surveillance and changes in management policy for white-tailed deer throughout the state.