2.9 | Population structure of B. schroederi in Qinling and Sichuan
A total of 240 samples collected from individuals in captivity and 26 samples from individuals in wild were re-sequenced using the DNBSEQ-T1&T5 platform. High-quality reads were aligned to the reference genome using BWA-MEM (0.7.13-r1126) (H. Li & Durbin, 2009) with default parameters. SAMtools (v0.1.19) (H. Li et al., 2009) and Genome Analysis Toolkit (GATK v 4.0.3.0) (Depristo, Banks, Poplin, Garimella, & Daly, 2011) were used to obtain the SNP set within the population. Hard filtering was applied to the raw variant set using ”QD < 2.0 || FS > 60.0 || MQ < 40.0 || MQRankSum < -12.5 || ReadPosRankSum < -8.0” –filter-name ”snp_filter”. SNPs with >0.5% missing data or <0.01 minor allele frequency (MAF) were filtered out using vcftools (v0.1.12a) (Danecek et al., 2011). PCA analysis of SNPs was carried out using EIGENSOFT (Nick et al., 2006) software, and the population clustering analysis was conducted in PLINK (Purcell et al., 2007). We used the whole-genome SNPs to construct the ML phylogenetic tree with 1000 bootstrap using iqtree (v1.6.12) (Lam-Tung, Schmidt, Arndt, & Quang, 2015), and using an genome sequence information ofP. univalens as an outgroup. Population structure of all was analyzed using the ADMIXTURE (v1.3.0) program with a block-relaxation algorithm. To explore the convergence of individuals, we predefined the number of genetic clusters K from 2 to 5.