Testing selection in the context of population genetics

Constant innovations in the eld of experimental evolution demand the development of adapted statistical test methods. Distinguishing between the processes of genetic drift and selection is an active eld of research in population genetics with new statistical approaches being published frequently. The implementation of genetic models as statistical test procedures is challenging due to the complexity of evolution. Newly published approaches are mostly compared to each other but only seldom to classic statistical tests. My simulations show that the Sign or CMH test outperform test procedures with sophisticated mathematical background explicitly developed for E&R experiments both in performance and runtime. Also, the most recent test method CLEAR (Iranmehr et al., 2017) gives comparable, but not superior, results to the classical test methods. In consequence, resequencing at intermediate generations is an unnecessary expense as long as new approaches do not outperform classical tests.submitted by Katharina SallingerUniversität Linz, Masterarbeit, 2017(VLID)227070