Figure legend
Figure 1 (a) Map of sampling locations. Annual sea surface temperature is indicated. (b) Genome-wide distribution of nucleotide diversity in 40 kb non-overlapping windows.
(c) Admixture analysis of the five Sillago japonica populations. The length of each colored segment represents the proportion of the individual genome inferred from ancestral populations (K = 2–6). (d) Principal components 1 and 2 for the 49 Sillago japonicaindividuals. (e) Neighbor-Joining tree constructed using p-distances among the 49 Sillago japonica individuals.
Figure 2 Plot of pairwise estimates of F ST/ (1 −F ST) versus two types of geographic distances (coastal distance and oceanic distance) between the populations.
Figure 3 Demographic history for each population inferred from PSMC analysis.
Figure 4 Pattern of population splits and mixture between the fiveSillago japonica populations. The drift parameter is proportional to Ne generations, where Ne is the effective population size. Scale bar shows the average standard error of the estimated entries in the sample covariance matrix.
Figure 5 Genomic regions with strong selective signals in populations ofSillago japonica. (a) Distribution of log2(θ πratios) andF ST values calculated in 40 kb sliding windows with 20 kb increments between the RS/ZS populations (ZS as control group). The data points in red (corresponding to the top 5% of empirical log2[θ π ratios] ratio distribution with values of >0.1204 and the top 5% of F ST distribution with values of >0.0904) are genomic regions under selection in the RS population. (b) Overlap candidate genes of the RS/ZS and RS/ST pairs based on a Venn diagram. (c) Overlap candidate genes of the ST/ZS and ST/RS pairs based on a Venn diagram. (d) Overlap candidate genes of the ZS/RS and Japan/RS pairs based on a Venn diagram. (e) Allele frequency of one SNP within the cold-temperature adaptation genePicalm  across the five S. japonica populations. (f) Allele frequency of one SNP within the high-temperature adaptation geneARHGAP42  across the five S. japonica populations.  (g) Allele frequencies of one SNP within the warm-temperature adaptation gene SORCS3 across the five S. japonica populations.
Figure 6 PCA based on the SNPs located in the candidate genes (a, cold-temperature adaptation genes; b, high-temperature adaptation genes; c, warm-temperature adaptation genes).
Figure 7 Top 20 KEGG pathway enrichment statistics for the candidate genes.
Figure 8 The predicted potential distribution (a, b), changes in habitat suitability (c) of China group under RCP45 scenarios and (d) response curves of predicted occurrence probability of China group against temperature.
Figure 9 The predicted potential distribution (a, b), changes in habitat suitability (c) of Japan group under RCP45 scenarios and (d) response curves of predicted occurrence probability of Japan group against temperature.