Genetic architecture of ideotype-related traits in middle American beans through single trait, multi-trait and epistatic genome-wide analyses

Abstract

Common bean is one of the major legume crops for direct human consumption. The genetic improvement of common bean is a primary approach to increase crop adaptability to climate change conditions while maintaining productivity. A diverse panel of middle American beans was evaluated over three growing seasons (2021, 2022 and 2023) for agronomic traits considered in the crop ideotype such as flowering, maturity, pigment content, lodging, and yield. A study of the genetic regions controlling trait variation was carried out using single-trait and simultaneous (joint) multi-trait GWAS approaches. Additionally, genome-wide epistatic interactions were also analyzed. Several previously reported and novel regions were identified as significant for individual traits in the single and multi-year analyses with varying percentages of individual (7-52%) and collective (10 - 59%) phenotypic variance explained. In the single-trait and multi-year analyses, markers detected for lodging showed the highest average of percentage of variance explained (52%) followed by other traits with percentages between 11 and 19%. For yield, new loci were found with estimated effects between -96.19 to 90.96 kg/Ha in the multi-year data. In the multi-trait analyses, marginal loci on Pv02 and another on Pv04 were identified to have interaction effects on flowering and yield. A significant locus on Pv04 showed a common effect between lodging and maturity, characterized by several SNPs. Significant epistatic interactions were found along different chromosomes for all the evaluated traits, with some loci having interactions with multiple regions. In flowering, an interaction between loci on Pv01 and Pv04 explained up to 10.5% of the phenotypic variation, followed by interactions between Pv06 and Pv10 for chlorophyll b and between Pv03 and Pv08 for yield explaining around 6% of the trait variation. Multiple transcription factors were identified as candidate genes, particularly in the pairs of combinations of epistatic effects. Based on the homology analyses of the candidate genes, several showed potential roles in the genetic control of the agronomic traits, particularly for flowering, maturity, and yield. Our results demonstrate the applicability of various approaches for common bean and show a comprehensive and expanded panorama of the genetic basis of agronomic traits. The results provide new resources for crop improvement that can be leveraged in multiple approaches such as selection, modeling and predicting crop performance and genetic gain.

Fig. 1

Phenotypic distribution of the BLUP-corrected means of the evaluated agronomic traits during different years and multi-year analysis. Broad-sense heritability (H²) for each trait is included in the top part of each panel

Fig. 2

Multi-trait GWAS for yield and DTF (A), and example of interaction effect observed for yield and DTF for the SNP with the highest effect (B). Analysis for DTM and lodging (C) and common effect between DTM and lodging for the most significant SNP (D). Significant associations are above the FDR corrected threshold (q < 0.05). REF denotes the reference allele and ALT is the alternative allele in the variant calling against the common bean reference genome v2.1. In B and D, black dots denote the mean and horizontal bars the median of the BLUP-corrected phenotypes for each allelic state combination

Fig. 3

Significant QTN-by-QTN interactions (QQIs) detected for the agronomic traits. The lines connecting different loci represent interactions for different agronomic traits coded as colors

Fig. 4

Example of epistasis for flowering time (A) and yield (B) with the interactions explaining the higher percentage of variance explained. REF denotes the reference allele and ALT is the alternative allele in the variant calling against the common bean reference genome v2.1. Black dots denote the mean and horizontal bars the median of the BLUP-corrected phenotypes for each allelic state combination