Contributions of climate to genetic divergence
Eight climate factors were included in the analysis of the contribution of climate to genetic divergence (Fig. 3C). A total of 272 SNPs were scanned by BayeScan and three GEA methods as significant outliers (Fig. 3D; Fig. S6). We observed a substantial intersection between the results of RDA and BayEnv regardless of whether the former accounted for population structure; 31 and 41 of the SNPs detected by BayEnv (i.e., 24% of each set) were detected by RDA.raw and RDA.correct, respectively. We also observed a sizable intersection (i.e., 33 SNPs; 20%) between the RDA.raw and RDA.correct sets. The 89 SNPs that were detected by at least two methods were considered putative adaptive SNPs (Fig. S3D). Forward selection also identified genetic structure (PC1-PC2) and BIO18 as the only significant predictors explaining the genetic allele frequencies of adaptive loci (Table S5). The 89 adaptive SNPs belonged to 34 of the 51 genes enriched in this study. Many of these are protein-coding genes that are known to be involved in the adaptation of other species. A complete list of genes, potential functions, and associated climate variables is given in Table S6.
Climate, demography, and geography together explained 71% of the variance by pRDA (Table 2; Fig. S6). When two of the three factors were controlled, demography was the only significant predictor, explaining 52.1% of the total variance, whereas pure climate and pure geography explained 1% and 2% of the total variance, respectively. These results suggest that the breaking of species boundaries by genetic groupings (e.g., Fig. 2B) due to hybridization may play a stronger role in enhancing the genetic diversity of these species within Taiwan Island than differentiation caused by geographic or climate barriers. Notably, 16.3% of the variance was confounded among the predictors (Table 2), suggesting some degree of collinearity among genetic structure, climate, and geography.