Dating back to Fisher, population genetics has a long standing tradition of interaction with statistics and probability. More recently, sequencing technologies lead to unprecedented amounts of data that provide ample opportunities to develop new and improved statistical methodology.

Motivated by challenges from applications, this talks discusses new and modified inferential approaches. More specifically, we talk about useful approaches to test for traces of selection in genomic time series data. We also discuss the estimation of the effective population size $N_e$ which is important to properly calibrate the null model of genetic drift.