GWAS-assisted and multi-trait genomic prediction for improvement of seed yield and canning quality traits in a black bean breeding panel

Authors: IZQUIERDO, PAULO - Michigan State University; WRIGHT, EVAN - Michigan State University; Cichy, Karen

Submitted to: G3, Genes/Genomes/Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/18/2024
Publication Date: N/A
Citation: N/A

Interpretive Summary: A major end use of dry beans in the U.S. is the canning market and new bean cultivars should be high yielding and have canning quality that meets industry standards. Simultaneously improving yield and canning quality is challenging due to antagonism of the two traits and specialized methodology needs for evaluating canning quality. In this study the usefulness of genomic prediction to improve genetic gain for seed yield and canning quality was evaluated. Genomic prediction accuracies were moderate for yield and canning appearance, and high for canned bean color retention and as genotypes from the new breeding cycle were included in the models, prediction accuracy tended to increase. Multi-trait models increased the prediction ability of complex traits, suggesting the feasibility of genomic prediction in early generations to increase selection intensity, and lead to higher genetic gain for seed yield and canning quality.

Technical Abstract: Black beans (Phaseolus vulgaris L.) are an important dry bean market class in the Americas, gaining popularity among U.S. consumers in recent years. Seed yield and canning quality are two key traits for new black bean cultivars. Simultaneously achieving breeding gains for both attributes is often hampered by negative trait associations. Furthermore, canning quality is challenging to evaluate in breeding programs because it requires specialized equipment and trained sensory panels. The integration of genomics and phenomics has the potential to increase selection accuracy and intensity in traits that are difficult to measure and have a complex genetic architecture such as yield and canning quality. Genomic prediction (GP) has shown significant potential in improving breeding progress for complex traits, and multi-trait models can boost prediction accuracies when using phenotypic correlated traits in GP. In this study, we evaluated the prediction accuracy of single-trait and multi-trait GP models, and the use of NIRS (Near-infrared spectroscopy) phenomics data on whole seeds, significant markers identified through genome-wide association (GWAS) in black bean advanced breeding lines over two breeding cycles. Our results showed that prediction accuracies were moderate for yield and canning appearance, and high for color retention when individuals from the same breeding cycle were used, and no significant differences were observed between single-trait and multi-trait models. However, when predictions were evaluated between breeding cycles, multi-trait models outperformed single-trait models by up to 41% and 63% for canning appearance and seed yield, respectively. The addition of significant SNP markers identified by GWAS and NIRS to estimate regularized selection indices using intact seeds in the models reduced the prediction accuracy within and between breeding cycles. As genotypes from the new breeding cycle were included in the models, prediction accuracy tended to increase. Our results demonstrated the potential of multi-trait models to increase the prediction ability of complex traits such as seed yield and canning quality traits in dry beans and highlighted the need of updating the training data set for the implementation of GP in a dry bean breeding program.