Institution

Abstract:

The biggest challenge when undertaking a quest for traces of selective events in molecular variation data is the stochastic noise arising from the interplay between genetic drift and the demographic events that the studied populations underwent. Among those demographic events, fluctuations in population sizes due to founder events (i.e. bottlenecks) can potentially create an important variance in the time to the most recent common ancestor of a sample of genes along a chromosome. When this is the case, it becomes impossible to simply rely on the assumption that selection leaves local genetic footprints whereas demography affects the whole genome. Indeed, demography does affect genetic variation along the whole genome, but in a stochastic manner which means that it will create several local deviations to neutrality that might then be misinterpreted as signatures of selection. Because of this serious problem, most of the selection-detection methods rely somehow on the users knowledge of the demographical history of his species. This project aims to provide an accurate estimation of the demography of three Drosophila melanogaster worldwide populations. This will be done by conducting an approximate bayesian computation analysis (ABC). The results of this projects will be an essential pre-requisite for projects aiming at detecting footprints of natural selection in genetic polymorphism variation data.