3.3 Population structure at different scales
M. alternatus population structure was compared at intermediate and fine scales. As shown in the phylogenetic tree, the population in the two regions was divided into two branches at the intermediate scale (Figure 2A). Principal component analysis confirmed this result, having the populations from the two regions clearly separated (Figure 2B). ADMIXTURE was also used to analyze the population genetic structure between the two regions. In the ADMIXTURE figure (Figure 2C), each column represents an individual, with the length of the different colored segments representing the proportion of an ancestor in the individual genome. Because the CV error was the smallest at K = 2 (CV error = 0.42066), the population genetic structure revealed two main differentiated gene pools (Figure 2C). Thus, the population in Shunchang was originated from one ancestor and the one from Xiapu was originated from a different ancestor, which is a conclusion consistent with the results of the phylogenetic tree and the PCA. In conclusion, the populations were clearly differentiated at the intermediate scale (>250 km), which clearly indicated the importance of geographic barriers against natural dispersal and gene flow of M. alternatus over long distances.
To reveal whether there were differences in population genetic structure at the fine scale, ADMIXTURE was also used to analyze the population genetic structures within Shunchang and Xiapu. According to the smallest CV error in Shunchang and Xiapu at K = 2 (CV error = 0.49264 and 0.47180, respectively; Supporting information Figure S1), the populations at the fine scale formed two clusters. Thus, the populations in SC included two clusters, and those XP included two other clusters (Figure 3). Although the fine-scale genetic structure detected by ADMIXTURE was weak, with an increase in the K value, seven meaningful clusters were obtained (Supporting information Figure S2).