Early detection of population fragmentation using linkage disequilibrium estimation of effective population size

Population subdivision due to habitat loss and modification, exploitation of wild populations and altered spatial population dynamics is of increasing concern in nature. Detecting population fragmentation is therefore crucial for conservation management. Using computer simulations, we show that a single sample estimator of N_e based on linkage disequilibrium is a highly sensitive and promising indicator of recent population fragmentation and bottlenecks, even with some continued gene flow. For example, fragmentation of a panmictic population of N_e = 1,000 into demes of N_e = 100 can be detected with high probability after a single generation when estimates from this method are compared to prefragmentation estimates, given data for ~20 microsatellite loci in samples of 50 individuals. We consider a range of loci (10-40) and individuals (25-100) typical of current studies of natural populations and show that increasing the number of loci gives nearly the same increase in precision as increasing the number of individuals sampled. We also evaluated effects of incomplete fragmentation and found this N_e-reduction signal is still apparent in the presence of considerable migration (m ~ 0.10-0.25). Single-sample genetic estimates of N_e thus show considerable promise for early detection of population fragmentation and decline.