Three main aspects of chicken biodiversity are dealt within this report: (a) cluster analysis based on autosomal microsatellites, (b) microsatellites on the sex chromosomes, and (c) SNP-based biodiversity.
(a) Cluster analysis of autosomal microsatellites: We used 29 microsatellites to genotype 2000 chickens randomly selected from 65 different populations representing various chicken types and various geographical regions. The computer program Structure placed the 65 populations into clusters that are in agreement with their geographic origin and breed history. Only at two predefined clusters, there is little admixture between non-commercial populations originating from Asia and those from Europe. In contrast, commercial broilers and brown egg layers appeared as admixed populations of these two main gene pools. Increasing the number of clusters resulted in generation of specific clusters of commercial lines, having very low admixture with other clusters. In addition, we identified seven mixed populations, each of which shared portions of their genome with several other genetic clusters.
(b) Microsatellites on the sex chromosomes: We predicted 173 potential microsatellites on chromosome W by in-silico analysis of the chicken genome assembly (version WASHUC1). Twenty five microsatellites of the highest sequence quality were tested in the lab for gender specificity. Unexpectedly, PCR products were generated in both sexes. Moreover, 14 selected microsatellites were mapped (using the East Lansing reference panel) and in all cases, the “W specific” microsatellites were mapped to chromosome Z and except for one locus, to the same ∼6 cM region. We conclude that the draft assembly for chromosome W is quite inaccurate.
(c) SNP-based biodiversity: Ten distinct chicken breeds were genotyped at 145 single nucleotide polymorphisms (SNPs) located at 14 random DNA fragments and twenty five, each from different and unlinked genes. Microsatellite genotypes of the same ten breeds were used for comparison. Applying bootstrap values as the criterion for tree's reliability, we found that: (1) increasing the number of SNPs had a higher impact on the reliability of the analysis than increasing the number of individuals per population, and (2) the bootstrap values of phylogenetic un-rooted trees based on microsatellites were relatively low.