| Abstract | Populations are often divided into subpopulations. Biologists use various statistics formed from genetic data to infer migration rates between subpopulations. Inference with such genetic statistics is difficult because little is known about their sampling distributions. Further, what little is known is tied to some specific model of migration patterns between the subpopulations. The statistic Fst serves as an example of this situation. The sampling distribution of Fst is not known and most of what is known about the distribution comes from simulation in which an island model or stepping stone model is assumed. In this talk we consider genetic statistics under a certain class of evolutionary models that we refer to as G/KC models. We show that in a large population limit, the island and 2-d stepping stone models are special cases of G/KC models. We then study the behavior of Fst under an arbitrary G/KC model and derive formulas that describe the distribution of Fst in the large sample setting. In this way we are able to study Fst in a general setting and free our analysis from a specific migration model such as the island model. Our approach is general and can be applied to other statistics such as heterozygosity measures. Our analysis uses coalescent based methods. |
|---|
