*2.3. Sequence Alignment and SNP Genotyping*

Stacks (version 2.55, Julian Catchen and Nicolas Rochette, Urbana, IL, USA) software is widely used in RAD-seq data analysis [12]. First, raw reads were processed using the process\_radtags program [11]. Through raw data processing, each clean read was assigned to a sample to ensure the quality of the sequencing data for use in subsequent analysis. Ustacks begins by clustering reads from a single sample obtained from the process\_radtags program to produce stacks [13]. Ustacks in the Stacks package was applied to cluster the reads for each sample, with the same stack representing one enzyme cutting site (locus). The clustering parameter -m was set to 3 [14], and the loci and locus sequencing depth of each sample were statistically analyzed. Next, we used Cstacks to merge the loci of all samples and allowed up to 2 mismatches between different sample loci to obtain the catalogue consensus sequence of each locus [14]. Sstacks was used next. The minor allele frequency (MAF) was set at 0.05 [15]. Loci in each sample deviating from the catalogue consensus sequence were identified, and then populations were filtered to obtain the SNPs. Only the genotype of the first SNP for each tag and loci shared by more than 75% of individuals were used for further analysis. The loci not in Hardy–Weinberg equilibrium (HWE) were removed [16], and the selected loci were used to analyze genetic diversity and population genetic differentiation.
