Independent estimates of population history help unlock the genetic signatures in duck populations

Background/Question/Methods

Genetic signatures can reveal species’ evolutionary and demographic pasts influenced by ancient environmental and geologic changes. The utility of DNA for making these inferences depends on the relationship between effective population size, mutation, and genetic drift (and in some cases immigration and emigration). More mutations and less genetic drift lead to higher levels of genetic diversity in larger populations than in smaller populations resulting in a linear positive relationship between effective population size and genetic diversity. However, the diversity of DNA polymorphisms can also be influenced by natural selection, complicating relationships between effective population size, mutation and drift; natural selection is more efficient in larger populations and can have different effects depending on the type of selection (balancing vs. stabilizing and directional). Disentangling the influences of selection and demographic histories on genetic structure requires additional data independent of the genome. We investigate the importance of demographic history and natural selection on genetic signatures in populations of Anas ducks around the world using a natural experiment. In this experiment, differing climate histories of five continents are the treatments, and 23 species serve as replicates (4-6 per continent). We estimate effective population sizes and long-term changes in demographic histories (i.e., growth rates) using coalescent models of six independent loci. We estimate population histories independently using species distribution models, backcasting based on palaeoclimate reconstructions.

Results/Conclusions

Estimates of population histories over a variety of climate histories concur between genetic models (GM) and distribution models (SDM) across the 23 species. First, effective population sizes from GM’s are significantly correlated with population sizes from census data. Second, growth rates from GM’s are significantly correlated with predicted changes from SDM’s. These results suggest that DNA polymorphisms are accurately reflecting demographic histories, and are consistent with selective neutrality for those loci. However, upon more detailed analysis, the residuals from the regression of GM growth rates on SDM predictions correlate with independent estimates of current population sizes. Because selection is more powerful in larger populations, this result suggests that selection has had a significant influence on genetic diversity. A more detailed analysis of genetic patterns among additional loci under varying strengths of selection will allow us to tease out the effects of selection to obtain more robust estimates of species’ demographic histories.