UC Berkeley

Improvements in sequencing technologies and the resulting increased availability of genetic data call for new and more sophisticated analysis methods. Particularly in ecological genetics and evolutionary biology, questions that can be addressed are often limited by the availability of analysis tools and statistical inference procedures. Non-equilibrium models in particular have been relatively poorly studied, mainly because analytical approaches are challenging and many useful and well-known results make equilibrium assumptions. However, using heuristic methods and strongly simplified models, we can make progress and arrive at procedures that help us gaining new insights from population genetic data. After an introductionary first chapter, in the second chapter, I develop an Approximate Bayesian Computation procedure to distinguish selection from standing variation from selection on a \textit{de novo} mutation. This method is applied to human genetic data where we identify two genes, ASPM and PSCA, that are most likely affected by selection on standing variation. In the third chapter, I develop an inference procedure to infer the origin of a range expansion, introducing the directionality index statistic psi. Applying this method to human data, we find a most likely origin of humanity in southern Africa, and evidence of the main expansion routes into Asia, finding evidence for a Southern route. In the fourth chapter, I extend the work on range expansions by developing an analytical model based on branching processes, which gives a biological interpretation to psi, and allows us to measure the decay of genetic diversity with distance. An application to Arabidopsis thaliana reveals that we are able to infer both recent expansions in the Americas, as well as expansions from the last glacial maximum in Europe. Between these three chapters, I use different approximation procedures to introduce inference procedures for models where direct likelihood calculations are not available.