Adaptive introgression

Fst outliers are often used on SNP data to detect selection (Wright 1951; Weir and Cockerham 1984; Excoffier et al. 2009), whilst the D-statistics (and its variants) is a frequently used method for detecting ancient admixture events (Green et al. 2010). We used the intersection of Fst outliers and genomic loci with significantly high D-statistic to identify introgressed regions from domestic ferret under selection in English polecats. We took the most likely trio of populations showing introgression (identified from our previous analyses) and filtered the input for scaffolds longer than 1 Mb where all samples had called genotypes. We then used Dsuite Dinvestigate to identify sliding windows of 1,000 SNPS and their associated admixture statistics (D, f_d, f_dM, and df) and selected the windows with the top 1-percent of f_dM values (Malinsky et al. 2020). The f_dM statistic (a modified version of the f_d statistic) is symmetrically distributed around zero under the null hypothesis of no introgression and can equally quantify shared variation between P3 and P2 (positive values) or between P3 and P1 (negative values).

Using VCFTools, we then identified the location of the top 1-percentile Fst outliers between the 2 “parental” populations of the introgressed trio (Danecek et al. 2011). We filtered down the Fst outliers to those that overlapped with the 1-percentile f_dM windows and then identified which of these intersections overlapped with genes in the domestic ferret genome annotation (Peng et al. 2014).