Tests of two methods for identifying founder effects in metapopulations reveal substantial type II error

Genetic analysis has been promoted as a way to reconstruct recent historical dynamics ("historical demography") by screening for signatures of events, such as bottlenecks, that disrupt equilibrium patterns of variation. Such analyses might also identify "metapopulation" processes like extinction and recolonization or source-sink dynamics, but this potential remains largely unrealized. Here we use simulations to test the ability of two currently used strategies to distinguish between a set of interconnected subpopulations (demes) that have undergone bottlenecks or extinction and recolonization events (metapopulation dynamics) from a set of static demes. The first strategy, decomposed pairwise regression, provides a holistic test for heterogeneity among demes in their patterns of isolation-by-distance. This method suffered from a type II error rate of 59-100 %, depending on parameter conditions. The second strategy tests for deviations from mutation-drift equilibrium on a deme-by-deme basis to identify sites likely to have experienced recent bottlenecks or founder effects. Although bottleneck tests have good statistical power for single populations with recent population declines, their validity in structured populations has been called into question, and they have not been tested in a metapopulation context with immigration (or colonization) and population recovery. Our simulations of hypothetical metapopulations show that population recovery can rapidly eliminate the statistical signature of a bottleneck, and that moderate levels of gene flow can generate a false signal of recent population growth for demes in equilibrium. Although we did not cover all possible metapopulation scenarios, the performance of the tests was disappointing. Our results indicate that these methods might often fail to identify population bottlenecks and founder effects if population recovery and/or gene flow are influential demographic features of the study system.