Genetic variation and population structure of Oryza sativa accessions in the AfricaRice collection and development of the AfricaRice O. sativa Core Collection

Genetic diversity studies provide an increased understanding of genebank collections and enhance their use for improving global food security. This study explores the genetic variation and population structure of 5738 rice (Oryza sativa L.) accessions representing 39.6% of the O. sativa collection conserved in the AfricaRice genebank of which 74.0% originated from African countries. These accessions were genotyped with 25,904 polymorphic DArTseq-based single nucleotide polymorphisms (SNPs). Genetic distances between pairs of accessions indicate high variability, with 21.0% of pairs being moderately distant and 78.2% highly distant from each other. The genotyped accessions are traditionally grown in six different agro-ecologies from 73 countries. Using neighbor-joining tree, principal component, and model-based population structure analyses, the accessions were divided into four genotypic groups representing the two O. sativa subspecies, Japonica (787 accessions) and Indica, which were further divided into landraces (1879 accessions), and improved cultivars (3027 accessions), and a fourth small group of admixed accessions. Subclusters identifying a specific agro-ecology (upland, lowland, and mangrove swamp) or originating country were noted. Using the maximum length sub-tree method, we selected 10% of the total accessions to form the "AfricaRice O. sativa Core Collection" (AROSCC). The subset of 600 O. sativa accessions captures more than 95% of the SNP polymorphisms in the entire collection. The AROSCC is an important resource to support pre-breeding and rice improvement programs around the world.