Neutral genetic variability and SNPs under selection
The analysis of genetic structure with Admixture yielded a single more plausible model that grouped all samples into a single cluster (K = 1, cv-error = 0.382), indicating no apparent genetic structure between both populations. Even when forcing the model towards higher values of K (i.e. less plausible results), we still failed to find any signs of between-populations genetic structure (K = 2, cv-error = 0.405; Figure 2).
Regarding outliers analyses, Bayescan did not detect any locus subject to selection (q-values range: 0.883 - 0.910). The largest increase in α, the magnitude of the locus-specific effect of selection, concentrated on four maximum (although low) values, two of which could not be annotated at all (Table 3). Of the other two, one could be annotated with a marginally non-significant E-value of 0.018 (conventional significance value = 0.01), which corresponded to the coding region of a repeated transmembrane ankyrin involved in the cellular response to salicylic acid (inflammatory response) and the regulation of the immune system. The other sequence was annotated with a much higher E-value (0.6) as the coding region of a nuclear photoreceptor protein that regulates gene expression based on the reception of steroids (described mutations point to impairments in the bearer’s vision). None of these SNPs presented allelic frequencies that deviated significantly from those expected under H-W conditions after applying the Bonferroni sequential correction (Table 3).