Population genetic structure and demographic analyses
The observed F ST values for COI, COII, and Cytb were 0.781, 0.687, and 0.803, respectively, indicating that regional populations are genetically isolated. For COI, the estimated migration rate (Nem, where Ne is the effective population size and m is the proportion of the population that migrates in each generation) was 0.07 migrants per generation (Slatkin, 1987; Slatkin & Barton, 1989). All three genes showed greater variation among regions (73.25%–75.90%) than within regions (0%–4.37%) (Table 3 and S4). We detected highF ST in pairwise combinations between regions E, F, G, and H (Table 4, S5, and S6). For the COI gene, 23 out of 28 pairwise combinations showed significant differentiation, and the highest pairwise F ST was 0.91731 for the comparison between region C and region G (Table 4).
Neutrality and population expansion parameters for each gene are summarized in Tables 5, S7, and S8. For COI, we detected negative Tajima’s D values for regions A, C, and D, indicating that the current haplotype diversity resulted from selection on certain genotypes. Tajima’s D for regions B, E, F, and H was not statistically significant, indicating neutral evolution. The τ values that represent the estimated time of expansion were very low in regions A, B, C and D (min = 0.0 in region B and max = 1.6 in region D), indicating sudden and recent population growth (Table 5). The τ values in regions E, F, and H were comparatively high (min = 9.2 in region E and max = 46.2 in region F), indicating that population growth was slower than that in regions A–D. The observed mismatch distribution was used to evaluate the demographic expansion history. The raggedness indexes for all regions except region E were not significant, suggesting that the expansion model could not be rejected, except in region E (Table 5). The analysis of region G, i.e., the USA population, was not informative because the samples showed no haplotype variation.