Diversity and population pattern of nuclear SNPs
Illumina reads from 26 genomes were successfully mapped on the reference with mean coverage ranging from 18.9x (30VOJ_SW) to 34x (29DZ_SW). None of the assemblies had to be filtered out due to low coverage and/or low Phred scores.
Patterns of population structure obtained using nuclear SNPs differed from genealogies based on maternally inherited genomes (Figure 4). In contrast to mtDNA, the structure derived by PCA from nuclear SNPs did not recognize only two mutually exclusive clusters corresponding to SW and SE mitochondrial lineages but also a third admixed cluster containing hybrid individuals. The pattern fitted very well geographical sampling along the anticipated contact zone in a transect running from north-west (SW population 2HBC_HorBl_SW) to the south-east (SE population 39RAD_SE). The geographical outlier, sample NE125b_Kot_SW from north-east Germany, was positioned within the hybrid samples on the main axis, whilst being separated from the rest of the samples along the other axis. This pattern may reflect a large geographical distance from the Czech localities combined with possible occurrence of a contact zone in NE Germany (Figure 4, map Figure 3).
Bayesian analysis in Admixture found the highest support for the occurrence of two clusters (see Figure S4 for cross-validation values). Each of the two clusters corresponded to one of the mtDNA lineages, with samples from the centre of the contact zone and the German specimen Ne125b_Kot showing admixed origin (Figure 4d).
ML tree-building method yielded a similar pattern to PCA (Figure S5). It revealed the two parental populations as monophyletic clusters and the hybrids forming an intermediate paraphyletic assemblage (with the long-branched Ne125b_Kot sample clustering among the hybrids; Figure S5).
Measures of genetic diversity corresponded very well with the previous methods. When divided into three populations (SW, SE, and hybrids), the highest pairwise FST value was detected between the SW and SE clusters (0.59), compared to the lower values between hybrids and SW or SE clusters (0.21 and 0.36, respectively, Table S4). Similarly, hybrids showed approximately 2 fold increase in heterozygosity compared to either of the SW and SE clusters (Table S4), as expected for individuals admixed between two lineages.