Additional file 1: Table S1. Sequencing data. Table S2. Summary of sequencing data. Table S3. Estimates of expected heterozygosity and observed heterozygosity. Table S4. Four rounds of increasingly focused optimizations used in δaδi. Table S5. Maximum-likelihood parameter estimates obtained from the joint demographic inference analysis 2D and 3D. Table S6. Genomic regions identified as candidate divergent regions. Table S7. GO enrichment (Top 20) of iHS significant selection sites in PEc and PEz&PEa clades. Table S8. KEGG enrichment of XP_EHH significant selection sites in PEc and PEz&PEa clades. Figure S1. SNPdensity. Figure S2. Sequencing depth. Figure S3. Karyotyping. Figure S4. The mapping ratio and shared SNPs distribution. Figure S5. Shared identity-by-descent (IBD) region among PEa, PEc and PEz populations. Figure S6. The shared IBD and paired Fst values of the three populations. Figure S7. Observed SFS between PEc and PEz & PEa clades. Figure S8. 3D divergence model used in δaδi. Figure S9. 2D divergence model used in δaδi. Figure S10. Fst distribution of PEc vs PEa & PEz. Figure S11. The iHS score distribution of PEc, PEa and PEz populations. Figure S12. KEGG enrichment of iHS significant selection locis of PEc and PEa & PEz respectively. Figure S13. GO enrichment of XP_EHH analysis of PEa and PEc & PEz respectively. Figure S14. Network diagram of GO function enrichment obtained by XPEHH analysis of PEa and PEc & PEz clades. Figure S15. Detection of mutations in three drug-resistant sites of β-tubulin. Figure S16. The π distribution and tajima’D distribution in the pgp-3 regions. Figure S17. The π distribution and tajima’D distribution in the unc-38, nrf-6, glc-1 and cup-4 regions. Figure S18. The π distribution and tajima’D distribution in the CYP3A31, CYP4C1, mrp-1 and CYP4V2 regions.