not-yet-known not-yet-known not-yet-known unknown ddRAD Library A ddRAD library was generated for 87 individuals. This dataset consisted of 8,329 SNPs. These were subsequently filtered into those SNPs that had been genotyped in >80% of the 87 individuals, leaving a final dataset of 2,300 SNPs. Additionally, one individual was run in triplicate to check data consistency. Only a few discrepancies in the genotypes (0.26%) were detected, including mainly some missing genotypes and some allelic dropouts. Only a few instances of non-Mendelian inheritance (0.9%) were detected using the datasets from the known related families. The juveniles from families were subsequently discarded from further analysis, leaving us with a data set of 81 individuals. The sampled individuals were chosen to be grouped in the same five population clusters as in the microsatellite analysis and the genetic diversity indices are given in Table 2. Estimates of heterozygosity values were much lower compared to those obtained from microsatellite data while inbreeding coefficient levels were close to zero or negative for all populations. Pairwise FST values range from 0.054 (UK Captive - Greek Captive) to 0.24 (Greece/North Macedonia - UK Captive). The phylogeographic structure of Greek and Scottish populations was assessed also based on the SNP data. Principal Component Analysis (PCA) revealed two main superclusters (Figure 5). The UK Wild clustered with all Captive samples while the Greek/North Macedonia Wild was clustered separately. Hybrid individuals were in the intermediate. Further analysis of each of the two main superclusters was carried out, to investigate at a finer resolution the phylogeographic structure of the populations. Due to the low quality of DNA from the UK feather samples, only 16 wild individuals were successfully genotyped using the 2,300 SNPs. Based on this small sample set, very little phylogeographic structure was observed. There was evidence of two genetic lineages that occur across all individuals but with a slightly higher representation of one of them in the birds from East Lothian (data not shown). As far as the East-Wild populations, although there are genotypes from fewer wild Greek individuals (n=43) for the ddRAD data, the resolution of 2,300 SNPs is much higher for each of those individuals than the microsatellite data providing strong evidence for five genetic lineages (Figure 6). These clusters geographically line with the lineages being broadly classified as, Drama, Grevena, Kozani, North Macedonia and Thessaloniki. The most distinct were the birds from Drama. The Thessaloniki group showed the lowest levels of distinctiveness, with some individuals aligning closely with the North Macedonia group. Also, similar to the microsatellite data, the individuals from Kozani show marked diversity, with some individuals showing similarities to the Grevena and North Macedonia populations. This aligns with the geographic position of Kozani, between Grevena and the Macedonian border. The individuals from Xanthi, and those collected in Greece by the (Liukkonen-Anttila et al. , 2002) study also shows affiliation with the birds from North Macedonia (Figure 6). Further analysis was focused on the ability of the SNP dataset to differentiate hybrids and evaluate individuals from each population cluster. Based on the presence of two superclusters (West/East), we examined the assignment of the individuals in these two groups (Figure 7). Hybrid samples are indeed clearly detected while a minor hybridization level is present in other individuals, such as the Greek Captive. Analysis for the identification of an informative SNP marker panel, for allowing population assignment and identification of non-native stocks for captive rearing, was further performed. All 2,300 SNPs were scored for their informativeness to distinguish an individual from either the UK Wild or the Greek/North Macedonia Wild population cluster. The 15 top-scoring SNPs were selected; alleles and genotype frequencies are shown in Table S5 (Appendix). In order to test the identification strength, assignment tests were performed by splitting the dataset randomly in reference and test datasets with a 70/30 ratio. All the tested individuals were assigned successfully to their population of origin. The 15-SNP panel identified the origin (West/East) of the individuals accurately with a score of 100%.