3.3 Population genetic structure and differentiation
After LD filtering, 1,545,006 SNPs common to di- and triploids were selected and used for population structure analysis. The three methods including PCA, structure analysis and genetic distance analysis all showed the same patterns of population genetic structure (Figure 3 and 4, and Figure S3). For diploids, two clear genetic clusters were shaped and consisted of individuals from invasive and native populations, respectively. The genetic cluster owned by native individuals split into three subclusters and followed a pattern of geographical separation: individuals from the mid Yangzte River (SS and ZDL); individuals from the lower Yangzte River (TH and CL) and Huai River (RR and HZL) and individuals from the Songhua River (DQ and JPL). For triploids, although several genetic clusters were observed, the pattern was not related to the geographical origin, with each cluster containing individuals from different rivers. The invasive triploids were assigned into different clusters. Results of PCA for all individuals revealed that both PCA1 and PCA2 did not clearly separate diploids and triploids, indicating patterns of population structure were not linked to the ploidy level.
For diploids, highly significant genetic differentiation (F ST) was detected for all pairwise comparisons among the invasive population (LL) and three genetic subclusters from the native range (Table 2). The F ST values between LL and two source genetic subclusters (i.e., the mid Yangzte River, the lower Yangzte River and Huai River) were 0.1045 and 0.0926, respectively, which were both higher than that between the two subclusters (0.0449). AMOVA analysis revealed that genetic variation between regions (invasive and source), among populations within regions and within populations were all significant and accounted for 6.29%, 8.65% and 85.06% of the total variation (P < 0.01), respectively. For triploids, low F ST values ranging from 0.0073 to 0.0281 (P > 0.05) were detected for pairwise comparisons between invasive (LL and CBL) and source (SC and NX) populations and a relative higherF ST (0.0345) was obtained between the two source populations (P < 0.01). The variation between regions, among populations within regions and within populations accounted for -2.93%, 4.41% and 98.52%, respectively. Overall, theF ST values between invasive and source populations were 0.0866 in diploids and 0.0027 in triploids.