Genomic selection on a panel of elite Andean breeding lines of common bean (Phaseolus vulgaris L.)

Common bean breeding in the tropics has benefited from the use of marker assisted selection to accelerate the breeding cycle. It uses individual markers previously identified to be significantly associated with targeted QTLs. However, some limitations arise when identifying and quantifying a large number of minor effect loci governing quantitative traits through standard QTL analysis and marker tagging. Genomic selection (GS) offers a promising approach for breeding by using many markers as predictors of the performance of individuals in a population1,2. This alternative strategy does not provide direct information about the underlying loci and genes that control quantitative traits, but instead offers a valuable selection criterion to be implemented in a breeding scheme. Different GS studies have been recently reported in plant breeding, targeting complex traits for many species with interesting results. In this study we want to assess the prediction accuracy of different genomic selection models in a collection of Andean bean breeding lines from CIAT