2.4 Population genetics analyses
Within-population diversity was assessed by expected heterozygosity (HE), average self-coancestry (s), fixation index (FIS), allelic diversity (A), and mean number of alleles per locus (K) using Metapop2 (López-Cortegano, Pérez-Figueroa, & Caballero, 2019). To analyze the population structure at the intermediate scale, a phylogenetic tree between Shunchang and Xiapu was constructed through 100 calculations of bootstrap values based on the neighbor-joining method and a principal component analysis was conducted, using MEGA 7.0 and genome-wide complex trait analysis (GCTA), respectively (Sudhir, Glen, & Koichiro, 2016; Yang, Lee, Goddard, & Visscher, 2011). ADMIXTURE version 1.3.0 (Alexander, Novembre, & Lange, 2009) was used to analyze population genetic structure at the two scales. The k genetic clusters were tested, with k ranging from one to seven, and the best k was determined according to the smallest cross-validation (CV) error. The vcftools (Auton & Marcketta, 2015) was used to calculate the FST between all pairs of sampled populations by averaging the values less than 0 after 5,000 bp windows and filtering (Supporting information Table S3, S4).