Population genetic analyses
Population genetic divergence
The microsatellite marker global FST value was 0.031 across all populations. The pairwise FSTvalues across all fifteen comparisons ranged from -0.0001 to 0.100 (Table 3). The Fisher’s exact test for microsatellite allele frequency distributions showed a highly significant population effect (p ≤ 0.001) in all but one population pairwise comparison (Utqiagvik/Svalbard; uncorrected p = 0.011), although that comparison was significant prior to Bonferroni correction (Table 3). The microsatellite pairwise FST values also showed highly significant population differentiation in 13/15 population pairs (FST : 0.009-0.100, p ≤ 0.012) comparisons (Table 3). The population pairs Alert/Mitivik Island and Utqiagvik/Svalbard (FST : -0.0001 for both pairs, uncorrected p -values of 0.564 and 0.464, respectively) did not show significant population differentiation before or after the sequential Bonferroni correction (Table 3). Combined results from the Fisher’s exact test and FST estimation at neutral markers provide evidence of significant divergence between all population pairs, with the exception of Utqiagvik/Svalbard and Alert/Mitivik Island population pairs which exhibited weak isolation.
The SNP marker global FST value was 0.022 across all populations. The SNP marker pairwise FSTvalues across all fifteen comparisons ranged from 0.004 to 0.053 (Table 4). The Fisher’s exact test for functional SNP marker allele frequency distribution showed highly significant population differentiation in 9/15 population pairs (p ≤ 0.009), with non-significant differentiation for: Aleutian Island/Pribilof Island, Alert/Utqiagvik, Alert/Mitivik Island, Alert/Svalbard, Utqiagvik/Mitivik Island, and Mitivik Island/Svalbard (uncorrected p -values: 0.56-0.88; Table 4). The SNP marker pairwise FST estimates matched the Fisher’s exact test results as the same population pairs (listed above) showed significant divergence (FST : 0.024-0.053, p ≤ 0.003; Table 4) and the remaining pairs showed non-significant divergence (FST : 0.004-0.013,p ≥ 0.039; Table 4). Broadly, the combined SNP Fisher’s exact test and FST estimation did not show significant genetic differentiation between the non-migratory populations (Aleutian and Pribilof islands), and among a majority (exception: Utqiagvik/Svalbard comparison) of the migratory populations (Alert, Utqiagvik, Mitivik Island, and Svalbard). However, all migratory - non-migratory population comparisons did show significant differentiation with the SNP marker data. The Utqiagvik/Svalbard and Alert/Mitivik Island population pairs had significant levels of SNP marker differentiation, although they were not significantly divergent based on neutral (microsatellite) marker data. Overall, our analyses show substantial genetic divergence among our six sampled populations.
Neighbour-joining cluster diagrams
The neighbour-joining (NJ) cluster diagrams (Figure 2) based on microsatellite and SNP genotypes show similar overall patterns of divergence. The microsatellite marker data show three strongly supported clusters (100% branch support): Utqiagvik & Svalbard, Alert & Mitivik Island, and Aleutian & Pribilof islands; although Aleutian and Pribilof islands also show strongly supported divergence (100% branch support) from each other (Figure 2a). On the other hand, the SNP genotype data do not show as strong support for population clustering among the six populations. However, the Aleutian & Pribilof islands, and Utqiagvik & Alert population pairs show strong patterns of divergence (99% and 100% branch support) between these populations and the others based on SNP marker data (Figure 2b).
Selection signatures at SNP loci