ddRAD statistical analysis
A genotype file was created containing the SNPs genotypes, called from the ddRAD data, for 87 individuals and 8,236 SNPs. All the SNPs were checked and were biallelic, and we proceeded with the filtering of the dataset. The PLINK toolbox (Purcell et al. , 2007) was used to remove loci that were not genotyped in ≥80% of the individuals. Additionally, a dataset containing three full-sib families (parents and two offsprings) was used to check the mendelian inheritance of the alleles.
The genetic diversity indices as well as a Principal Component Analysis (PCA), based on Nei’s genetic distance (Nei, Tajima and Tateno, 1983) were calculated using GENEALEX 6.502 software (Peakall and Smouse, 2006). Population pairwise FST was calculated using Arlequin 3.5 (Excoffier and Lischer, 2010). The genetic structure of grey partridge was investigated using STRUCTURE 2.3 (Pritchard, Stephens and Donnelly, 2000). The log likelihoods of our data set [ln Pr(X|K)] were estimated for different numbers of genetic clusters using an admixture ancestry model based on 100,000 burn-in steps followed by 1,000,000 Markov chain Monte Carlo (MCMC) replicates. We utilized a method developed by (Evanno, Regnaut and Goudet, 2005) to determine the number of populations present, based on the second-order rate of change in the log probability of the data (ΔK) among 20 runs of each assumed K using the web-based utility “Harvest” (http://taylor0.biology.ucla.edu/struct_harvest). We then evaluated the membership coefficient value (q) for each individual.
Analysis for the identification of an informative SNP marker panel for allowing population assignment and identification of non-native stocks for captive rearing was performed based on the ”Informativeness of assignment” method, as implemented in TRES software (Kavakiotis et al. , 2015). Furthermore, assignment tests were performed according to Rannala and Mountain method (Rannala and Mountain, 1997) in GENECLASS2 software (Piry et al. , 2004).