Genetic diversity and differentiation
The expected heterozygosity (H exp), observed heterozygosity (H obs),F IS, number of polymorphic sites, and πfor each population and species were estimated using Arlequin (Excoffier & Lischer, 2010). The number of private alleles was calculated by the package poppr (Kamvar, Tabima, & Grünwald, 2014) in R 3.6.1(R Core Team, 2013).
Next, to evaluate the genetic differentiation and ancestry of theScutellaria species, principal component analysis (PCA) was performed using adegenet (Jombart, 2008) in R 3.6.1 (R Core Team, 2013). All SNP sites were retained, and missing data were imputed with mean values for each SNP site. We further estimated individual assignments using StrAuto (Chhatre & Emerson, 2017) with 100,000 MCMC chains, 25,000 steps as burn-in, and ten iterations from K = 1 toK = 20. The results were uploaded to CLUMPAK to generate consensus plots and evaluate the optimal K using the ΔKmethod (Kopelman et al., 2015). Finally, the phylogenetic relationships among the species network were reconstructed using the NeighborNet program with \soutthe concatenated supermatrix of all retained genes as input implemented with SplitsTree 6 (Huson & Bryant, 2006).