Genomics of iron and zinc concentration, iron boavailability, and yield in common bean: multi-locus GWAS, structural equation modeling, candidate gene prioritization and genomic selection

Authors: AMONGI, WINNYFRED - Makerere University; NKALUBO, STANLEY - National Agricultural Research Laboratories; OCHWO-SSEMAKULA, MILDRED - Makerere University; ARFANG, BADJI - Regional Universities Forum For Capacity Building In Agriculture (RUFORUM); ODONG LAPAKA, THOMAS - Makerere University; NUWAMANYA, EPHRAIM - Makerere University; TUKAMUHABWE, PHINEAS - Makerere University; DRAMADRI ONZIGA, ISAAC - Makerere University; Cichy, Karen; MUKANKUSI, CLARE - National Agricultural Research Laboratories

Submitted to: Crop Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/18/2026
Publication Date: 3/25/2026
Citation: Amongi, W., Nkalubo, S., Ochwo-Ssemakula, M., Arfang, B., Odong Lapaka, T., Nuwamanya, E., Tukamuhabwe, P., Dramadri Onziga, I., Cichy, K.A., Mukankusi, C. 2026. Genomics of iron and zinc concentration, iron boavailability, and yield in common bean: multi-locus GWAS, structural equation modeling, candidate gene prioritization and genomic selection. Crop Science. 66. Article #e70262. https://doi.org/10.1002/csc2.70262.
DOI: https://doi.org/10.1002/csc2.70262

Interpretive Summary: Genetic improvement of consumer-related and nutritional quality attributes are important goals of dry bean breeding programs globally. Reducing cooking time and increasing micronutrient density and bioavailability are key consumer focused attributes of many breeding programs. Genetic improvements for these traits have largely been made through direct phenotypic measurements which are expensive and labor intensive. This is despite the availability of significant genomic resources that should allow more reliance on DNA related selection methods. The goal of this study was to evaluate genomic resources including genome wide associations, candidate genes, and genomic prediction for cooking time, micronutrient levels, and micronutrient bioavailability in dry bean breeding populations. Favorable associations, candidate genes, and genomic prediction models were identified. Zinc was the most accurately predicted trait, followed by iron. Environmental factors impacted genomic predictions. This study provides information on applying genomic tools to dry bean breeding programs focused on consumer related traits.

Technical Abstract: Common bean is a legume crop of global importance because of its nutritional composition and contribution to the ecosystem. The crop is a staple food for millions of people in Africa and serves as a vital tool for iron (Fe) biofortification. However, the production and consumption of the crop is hindered by various biotic and abiotic factors, long cooking time and the presence of bioactive compounds that limit bioavailable Fe (BioFe). Breeding has greatly relied on phenotypic selection to develop improved varieties for these traits. However, this method is time-consuming, environment dependent, and often results in marginal genetic gains. To accelerate genetic gain of multiple traits per unit time, the study identified underlying genetic basis of, iron (Fe), zinc (Zn), bioactive compounds, cooking time (COOKT), hydration coefficient (HC), seed coat color lightness or darkness and yield, including the number, frequency, effect sizes of genetic variants that contribute to the variation in target traits. Additionally, the study optimized genomic prediction models for the effective combination of yield, Fe, Zn, COOKT, HC, total polyphenols, flavonoids and phytate. Genotyping by sequencing (GBS) was used to generate single nucleotide polymorphic (SNP) markers in 427 bean genotypes. The quantitative trait loci (QTL) associated with target traits were determined using multi-locus models for single traits. To discriminate pleiotropic and single-trait SNPs, a multi-trait multi-locus structural equation modeling (SEM) was also performed. Forty-seven and 35 QTL on all the 11 bean chromosomes expressed significant desired effect on the traits by the two methods. Four QTL on Pv02, Pv05, and Pv11 colocalized for increased Fe and Zn. The traits Zn and HC, flavonoids and HC, yield and total polyphenols, COOKT and HC, and yield, flavonoid and Zn were favorably associated with the same genomic regions on Pv04, Pv05, Pv06, Pv07, and Pv07, respectively. Nine QTL were identified as pleiotropic in six traits, however, only two were favorable for all traits. The identified genomic regions included those previously reported and new regions. Over 100 genes were prioritized for different traits but the expression pathways of 11 genes including Phvul.004G007100 (flavonoids), Phvul.001G005200, Phvul.007G008500, Phvul.007G008600 and Phvul.009G134700 (cooking time), Phvul.002G008400 (cooking time and bioactive compounds), Phvul.002G182000 (yield, Fe, Zn), Phvul.002G063700 (Fe and Zn), and Phvul.001G265300, Phvul.007G170900, and Phvul.007G171800 (Zn) were related to traits in the parentheses. Incorporating major favorable QTL as fixed factors resulted in the highest prediction accuracies for all traits except flavonoids, with values ranging from 0.43 to 0.96. The beneficial markers and associations improved accuracy of genomic prediction.