Exploring the genetic architecture and improving genomic prediction accuracy for yield, mineral concentration, and canning quality traits in common bean (Phaseolus vulgaris)

Dry bean (Phaseolus vulgaris L.) is the most important legumes for human consumption worldwide and is an important source of protein, vitamins, and micronutrients in the human diet. This research aimed to i) uncover the genetic architecture of yield, Fe bioavailability and seed micronutrient concentration, ii) characterize the genetic control of canning quality traits, and ii) assess the accuracy of genomic prediction models for yield and end-use quality traits. The genetic architecture of yield and seed micronutrient concentration was assessed through a combination of meta-QTL analyses integrating published studies over the last two decades in dry bean. A Gaussian mixture model was used to determine the number of distinct QTL in the meta-QTL analyses. Consistent meta-QTL over different genetic backgrounds and environments were identified, reducing the confidence interval compared with initial QTL. Furthermore, a genome-wide association (GWA) study with 295 lines of the yellow bean collection and 82 yellow recombinant inbred lines identified a major QTL for Fe bioavailability related to the ground factor P gene. A black breeding panel with 415 lines was evaluated for yield and canning quality traits in two growing seasons. Consistent associations for color retention, appearance and texture of canned beans were identified across years. Genomic prediction models provided moderate to high accuracy for end-use quality traits on the yellow and black populations. The genomic prediction accuracy was related to the heritability of each trait, and improvement of accuracy was observed for complex traits when secondary traits were included in the model, while for traits with major QTL, the use of associated markers as fixed effects increased prediction ability. The use of meta-QTL analyses and GWA in this study lays a foundation of the genetic control of yield and end-use quality traits and reveals the potential of genomic prediction for these traits in dry beans.