Fig. 2

a Best-supported partial redundancy analysis (pRDA) for the E. blandingii target dataset, depicting the relationship between environmental axes and the first axis of genetic differentiation (PC1). Blue arrows indicate supported environmental relationships comprising redundancy axes; note that the best-supported pRDA included only a single predictor variable (road density) represented along the first constrained axis (RDA1). Colored points indicate loadings of samples from each sample site in multivariate space, and interposed red text indicates loadings of each identified marker for CIRBP, HSPA8 (HSPA8_1, HSPA8_2, HSPA8_3), and TRPV1 in multivariate space. b Variance partitioning between climate, land cover, and geography for the E. blandingii target dataset, indicating higher percent variance explained for the land cover dataset. Overlapping regions between circles indicate variance partitions correlated across variable classes. c-d best supported pRDA for the relationship between climate variables (c) or land cover variables (d) and genetic variation in the E. blandingii SNP dataset, with coloring of lines and symbols as described above. e Variance partitioning between climate, land cover, and geography for the E. blandingii SNP dataset