We also calculated the amount of differentiation between each population and all other populations, as the fixation index, F_ST, using the total polymorphism across the Muller elements (Drosophila chromosomes) (Weir and Cockerham 1984). F_ST appears to be generally low across the genes in theD. innubila genome (Figure 1B, total median = 0.00567), consistent with nuclear gene flow between populations (Weir and Cockerham 1984). In contrast, there is higher F_STbetween mitochondrial genomes (Figure 1C). Both nuclear and mitochondrial results are consistent with the Structure/StairwayPlot results. However, consistent with a more recent population contraction upon migration into PR, F_ST of the nuclear genome is significantly higher in PR (Figure 2A, GLM t-value = 93.728,p -value = 2.73e-102). Though PR F_ST is still extremely low genome-wide (PR median = 0.0105), with some outliers on Muller element B like other populations (Supplementary Figure 2). We also calculated the population genetic statistics pairwise diversity and Tajima’s D for each gene using total polymorphism (Tajima 1989). As expected with a recent population contraction in PR (suggesting recent migration and establishment in a new location), pairwise diversity is significantly lower (Figure 2B, GLM t-value = -19.728, p -value = 2.33e-86, Supplementary Table 2) and Tajima’s D is significantly higher than all other populations (Figure 2C, GLM t-value = 4.39, p -value = 1.15e-05, Supplementary Table 2). This suggests that there is also a deficit of polymorphism in general in PR, consistent with a more recent population bottleneck, removing rare alleles from the population (Figure 2C, Supplementary Figure 3). Conversely, the other populations show a genome wide negative Tajima’s D, consistent with a recent demographic expansion (Supplementary Figure 3).

Figure 2: Summary statistics for each population. A.Distribution of F_ST across genes for each population versus all other populations. B. Distribution of pairwise diversity for each population. C. Distribution of Tajima’s D for each population. D. F_ST distribution for Antifungal associated genes for each population. E.F_ST distribution for cuticular proteins for each population. F. F_ST distribution for all immune genes (excluding antifungal genes). In A, B & C all cases significant differences from CH are marked with an * and outliers are removed for ease of visualization. In D, E & F, significant differences from the genome background in each population are marked with a + and white diamond mark the whole genome average of F_ST for each population.