Genetic diversity and phylogeographic analysis
The sequences were checked visually and edited manually by BioEditor ver. 7.09 (Hall, 1999), and then aligned by Clustal W in MEGA ver. 5.0 (Tamura et al., 2011). The three cpDNA sequences were concatenated for subsequent analysis. Identification of cpDNA haplotypes (H) and ITS ribotypes (R), and calculation of their diversities (H d/R d) as well as nucleotide diversity (P i) were performed by DnaSP v. 5.0 (Librado & Rozas, 2009). Analysis of molecular variance (AMOVA) was conducted to estimate the genetic variances within and between populations and groups by Arlequin ver. 3.1 (Laurent et al. , 2005). Network of cpDNA haplotypes and ITS ribotypes were drew by TCS1.2.1 (Clement et al. , 2000). Phylogenetic analysis, takingPrunus salicina , Cerasus tomentosa and C. yedoensisas outgroups, was constructed based on maximum parsimony (MP), maximum likelihood (ML) and Bayesian(BI) to make the best use of the CIPRES Science Gateway V. 3.3. (Miller & Gross, 2011). The robustness of each internal branch of the genealogical trees was evaluated with 1,000 bootstrap replicates. All the above analysis involved both cpDNA and ITS sequences.
GST, representing haplotype frequency, and NST, representing haplotype difference, were calculated to examine the existence of population structure by PERMUT v. 2.0 with 10,000 permutations (Pons & Petit, 1996). N ST> G ST, P < 0.05 indicates the presence of a phylogeographic structure. Mismatch distribution analyses and neutrality tests were conducted to test population expansion using ARLEQUIN v. 3.5 with 10,000 permutations. These analysis were only conducted with cpDNA sequences.