A new $F_{\text{ST}}$-based method to uncover local adaptation using environmental variables
- Genome-scan methods are used for screening genome-wide patterns of DNA polymorphism to detect signatures of positive selection. There are two main types of methods: (i) 'outlier' detection methods based on Fst that detect loci with high differentiation compared to the rest of the genomes, and (ii) environmental association methods that test the association between allele frequencies and environmental variables. - We present a new Fst-based genome-scan method, BayeScEnv, which incorporates environmental information in the form of 'environmental differentiation'. It is based on the F-model, but, as opposed to existing approaches, it considers two locus-specific effects; one due to divergent selection, and another one due to various other processes different from local adaptation (e.g. range expansions, differences in mutation rates across loci or background selection). The method was developped in C++ and is avaible at http://github.com/devillemereuil/bayescenv. - Simulation studies shows that our method has a much lower false positive rate than an existing Fst-based method, BayeScan, under a wide range of demographic scenarios. Although it has lower power, it leads to a better compromise between power and false positive rate. - We apply our method to human and salmon datasets and show that it can be used successfully to study local adaptation. We discuss its scope and compare its mechanics to other existing methods.