Genome-Enabled Prediction Models for Yield Related Traits in Chickpea

Manish Roorkiwal, Abhishek Rathore, Roma R. Das, Muneendra K. Singh, Ankit Jain, Samineni Srinivasan, Pooran M. Gaur, Bharadwaj Chellapilla, Shailesh Tripathi, Yongle Li, John M. Hickey, Aaron Lorenz, Tim Sutton, Jose Crossa, Jean-Luc Jannink, Rajeev K. Varshney

Genomic selection (GS) unlike marker-assisted backcrossing (MABC) predicts breeding values of lines using genome-wide marker profiling and allows selection of lines prior to field-phenotyping, thereby shortening the breeding cycle. A collection of 320 elite breeding lines was selected and phenotyped extensively for yield and yield related traits at two different locations (Delhi and Patancheru, India) during the crop seasons 2011-12 and 2012-13 under rainfed and irrigated conditions. In parallel, these lines were also genotyped using DArTseq platform to generate genotyping data for 3000 polymorphic markers. Phenotyping and genotyping data were used with six statistical GS models to estimate the prediction accuracies. GS models were tested for four yield related traits viz. seed yield, 100 seed weight, days to 50% flowering and days to maturity. Prediction accuracy for the models tested varied from 0.138 (seed yield) to 0.912 (100 seed weight), whereas performance of models did not show any significant difference for estimating prediction accuracy within traits. Kinship matrix calculated using genotyping data reaffirmed existence of two different groups within selected lines. There was not much effect of population structure on prediction accuracy. In brief, present study establishes the necessary resources for deployment of GS in chickpea breeding.