Figure 1.
Selection history of the experimental populations.
The Grand Ancestor population (GA) was maintained in benign laboratory conditions, and was used to initiate populations maintained on salt-enriched media (AS) or cadmium-enriched media (AC). The treatment populations were produced by crossing two ancestral population AS and AC. There are 5 replicate populations of each of the four treatments (not illustrated).
Figure 2.
Mean survival in cadmium (A) or salt (B) environment.
Each point represents the mean survival of ∼140 full-sib families for each population. Each bar represents the mean across five replicate populations for each treatment.
Figure 3.
Differentiation between salt and cadmium populations.
Sliding-window (5 kb non-overlapping) plots of differentiation along chromosome 3R: the average -log(q-value) from the CMH test (A) and the average difference in mean allele frequency (B) between the salt and cadmium populations. For each environment, the data are calculated based on the environment-specific ancestor and the five replicate populations. The ratio of the number of genic to intergenic sites (C) and the ratio of the number of coding sites to intronic sites (D) in different -log(q-value) bins using data from the whole genome. To compare the magnitude of the ratios for different functional catergories, the ratios are standardized around 1 by dividing the mean ratio among the 11 bins. Strong significance corresponds to larger values of -log(q-value).
Table 1.
The number of highly differentiated sites between each pair of treatments.
Figure 4.
Average diversity (π) within experimental populations across the genome.
Each point represents the mean π across the common sliding windows (including non-variable sites) for each population. Though ancestral populations are shown for reference, we are primarily interested in the mean diversity (π) among the four treatments that were created by crossing AC and AS. There is significant variation among the four treatments (F3,16 = 9.68, P = 0.0007). The “b” group is significantly higher than the “a” group based on ANOVA Tukey HSD test. Spatial has significantly higher diversity than the Temp (Padj = 0.0006, Tukey HSD), the Salt (Padj = 0.026) and the Cad (Padj = 0.0036).
Figure 5.
Mean π within treatments as function of differentiation (d) between the ancestral source populations (AC and AS).
(A) and (B) The average π across different levels of ancestral differentiation for (A) all χ-site SNPs or (B) only those χ-site SNPs that have high initial diversity (πini>0.4). The x-axis is the allele frequency difference between ancestral populations, d = |pAC−pAS|. Error bars represent the standard error among the five replicates for each treatment. (C) and (D) Comparison among treatments in average π for sites that have high initial diversity (πini>0.4) using sites with (C) weak ancestral differentiation (d<0.3) or (D) strong ancestral differentiation (d>0.7). For the weakly differentiated sites (C), Spatial has significantly higher diversity than the Temp (Padj = 0.006, Tukey HSD) and the Salt (Padj = 0.028). For the highly differentiated sites (D), both Spatial and Temp treatments have significantly higher diversity than the two constant treatments (Padj<0.03 between any a & b pairs).
Figure 6.
FST values for each treatment in different regions.
We screened for the χ-sites above with πini>0.4 and divided them into five recombinationally independent regions of the genome. For weakly differentiated sites (d<0.3, Figure 6A), There is significant variation among treatments in FST values with Temp being highest among them (F3,16 = 6.72, P = 0.0038). The Temp has significantly higher FST than Spatial (Padj = 0.002). The difference between Temp and the two constant treatments are marginally non-significant (Temp vs. Salt: Padj = 0.084; Temp vs. Cad: Padj = 0.087). For highly differentiated sites (d>0.7, Figure 6B), the FST values do not significantly differ among treatments (F3,16 = 1.16, P = 0.35). For both weakly and highly differentiated sites, the treatments do not differ in expected total heterozygosity (HT) (F3,16 = 1.02, P = 0.41, Figure 6C; F3,16 = 1.59, P = 0.23, Figure 6D).