Two hundred forty-six snap bean genotypes and 49 dry beans representing both centers of domestication and six bean races with materials from Europe, Asia, and the Americas were genotyped using a single nucleotide polymorphism (SNP) array. The data was analyzed for expected heterozygosity, K-means clustering, principal components, phylogenetic relationships, and population substructure. When all gene pools of snap bean were assembled, the expected heterozygosity was roughly equivalent to a carefully chosen panel of dry beans representing all bean races and centers of domestication demonstrating the genetic richness of snap materials in total. K-means clustering and K = 2 structure analysis showed significant mixing of gene pools in the European and American commercial snap materials and the dominance of the Andean center of domestication among commercial contemporary snap beans. Conversely, the same analysis showed that Chinese, Iberian, and heirloom materials were underrepresented in contemporary materials. Further, Structure analysis revealed eight distinct groups within snap beans. Two showed strong kinship to the Middle American center of domestication, three to the Andean center of domestication, and three showed admixture between the two centers. Snap beans may have been independently derived from dry beans more than once and from both centers. Overall, we identified eight potential germplasm pools for snap bean.
Keywords: K-means clustering; Phaseolus vulgaris; Structure; common bean; domestication; expected heterozygosity; gene pools; green bean; phylogenetics; principal components; single nucleotide polymorphism.