Temperature-related natural selection in a wind-pollinated tree across regional and continental scales

Adaptive genetic variation is a key factor in evolutionary biology, but the detection of signatures of natural selection remains challenging in nonmodel organisms. We used a genome scan approach to detect signals of natural selection in the Black alder (Alnus glutinosa), a widespread wind-pollinated tree. Gene flow through pollen dispersal is believed to be high in this species, and we therefore expected to find a clear response to natural selection. In combination with two different landscape genetic approaches, we determined which environmental variables were most associated with the inferred selection. This analysis was performed on a regional scale (northern Belgium) and on a continental scale (Europe). Because climate-related differences are much more pronounced at the continental scale, we expected to find more selection-sensitive genetic markers across Europe than across northern Belgium. At both spatial scales, a substantial number of genetic loci were considered outliers, with respect to neutral expectations, and were therefore identified as selective. Based on results from our combined approach, four putative selective loci (or 2.5%) were recovered with high statistical support. Although these loci seemed to be associated with different environmental variables, they were mainly temperature-related. Our study demonstrates that the use of complementary methods in landscape genetics allows the discovery of selective loci which otherwise might stay hidden. In combination with a genome scan, the selective loci can be verified and the nature of the selection pressure can be identified.