Figure 2

GCNVs likely under parallel selection. The normalized numbers of mapped reads per 1-Mb gene length for each gene across the genomes of the (A) freshwater and (B) marine groups. Each black point represents the number for each gene in each individual, and the green lines represent the mean values for each gene across individuals. (C) The false discovery rate of the EdgeR analysis on the differences in the numbers of mapped reads between the freshwater and marine groups for each gene. Asterisks indicate the positions of the GCNVs under parallel selection (FDR < 0.05).