Outlier loci detection
The “all-species” and species-specific datasets analyzed in Genepop were also run through Bayescan v.2.1 to identify loci under divergent selection (Foll, 2012). Parameters of the Markov chain Monte Carlo algorithm were set to 20 pilot runs of 5000 iterations. Afterward, a burn-in of 50,000 iterations followed by 50,000 iterations were used for estimation with a thinning interval of 10. Jeffrey’s scale was used to interpret selection per loci (Jeffreys, 1998). Loci with a log10 value > 0.5 are considered to have “substantial” evidence of selection and those with a log10 value > 1.0 have “strong” evidence of selection. To identify loci under selection across clusters another new SNP dataset was generated by filtering to include only those occurring in all five clusters and 75% of the individuals within each cluster. The nucleotide sequences for each locus found to be under selection were submitted for an alignment search in the InsectBase and Flybase databases (Thurmond et al., 2019; Yin et al., 2016).