Analysis of genetic structure in a sample of coffee (Coffea arabica L.) using fluorescent SSR markers

The knowledge of population structure is important to determine the degree of linkage disequilibrium, which allows the selection of genotypes for association mapping. Using 47 SSR markers, the genetic variability and population structure of 68 accessions of C. arabica (wild and cultivated) and of three diploid species used as reference were evaluated. The analysis was done with the distance method and the structure model. The structure analysis inferred nine subpopulations (k = 9), for which the greatest values of probability were obtained. Three of the groups corresponded to the three diploid species as expected. There were six groups identified within C. arabica. The genetic subdivisions within C. arabica were based on geographical origin, degree of domestication, and dispersal history of coffee. One group consisted entirely of cultivated genotypes, where intense population bottleneck were associated with a founder effect. This was the most homogeneous group, as demonstrated by the reduced distance between cultivars in the dendrogram. Three of the cultivated genotypes, originating from Sudan, were separated into an independent group, presumably due to selective adaptation to a different set of environmental conditions. Another group consisted of genotypes of the type "ennarea" that were grown and cultivated in isolation on the shores of the Tana lake. The semi-wild genotypes clustered into three different groups. This type of analysis provides a strong evidence of population structure in C. arabica. Based on these findings, it is possible to better identify a balanced sample of diverse plants in germplasm.