A comprehensive phenotypic and genomic characterization of Ethiopian sorghum germplasm defines core collection and reveals rich genetic potential in adaptive traits.

ABSTRACT

Understanding population genetic structure and diversity of a crop is essential in designing selection strategies in plant breeding. About 2010 Ethiopian sorghum accessions were phenotyped for different traits at multiple locations. A subset of the collection, 1628 accessions, predominantly landraces, some improved varieties, and inbred lines were genotyped by sequencing. Phenotypic data revealed association of important traits with different sorghum growing agro-climatic regions, high genetic diversity and the presence of rare natural variation in the Ethiopian sorghum germplasm. Subsequent genotypic analysis determined optimum number of sub-populations, distinct cluster groups and ancestries of each sorghum accessions. To improve utilization of germplasm, a core subset of 387 lines were selected following posteriori grouping of genotypes based on cluster groups obtained through GBS analysis followed by stratified random sampling using quantitative traits. In order to evaluate how well this new sorghum and millet innovation lab (SMIL) collection from Ethiopia is represented within the largest world sorghum collection at United States Department of Agriculture - National Plant Germplasm System (USDA-NPGS) and the sorghum association panel (SAP), comparisons were conducted based on SNP data. The SMIL collection displayed high genetic diversity with some redundancy with the USDA-NPGS germplasm but SAP showed clear distinction. Furthermore, genome-environment association analysis identified candidate genes associated with adaptation to abiotic factors, that will be important for exploitation of adaptive potential to different environments. In summary, our results described the diversity and relationship of sorghum collections, representativeness of developed core and provide novel insights into candidate genes associated to abiotic stress tolerance.