B-1
Selection of Lima Bean (Phaseolus lunatus) Genotypes for Tolerance to Drought Stress
Carlos German Muñoz-Perea, Universidad Nacional de Colombia
carlos68munoz@hotmail.com
Co-authors: Carlos German Muñoz Perea, Yolanda Potosi, María Isabel Chacón Sanchez
Lima bean (Phaseolus lunatus) is a legume of great agricultural and nutritional importance in different world regions. It is cultivate for their edible seeds rich in protein, fiber and micronutrients. One of the major problems for its production is intermittent or terminal drought that it could be increasing for climate change in the next years. This research has the main goal to select lima bean genotypes that have good agronomic and yield performance under drought stress conditions that could be a alternative to food production in growing regions with water deficit and high temperatures. In the Experimental Research Center of the Universidad Nacional de Colombia at Candelaria (CEUNP), 104 lima bean genotypes from CIAT germplasm bank were planted each one in 3.0 m soil rows in three screen houses with impermeable roof under two environments, normal irrigation and drought stress in February of 2025. These genotypes were been previously selected in 2024 for drought tolerance in vegetative stage at Medellin and Bogota Campus. In the drought environment, irrigation was cut-off before genotypes flowering and it was not irrigated again until experiment was harvested. All genotypes showed tolerance to drought stress at vegetative stage however there are differences among genotypes at reproductive stage. Almost half of genotypes produced seeds. Eight genotypes showed outstanding seed production and agronomic performance under drought stress. These genotypes will be test proximately in a yield essay and will be determined water and nutrient use efficiency under irrigated and no irrigated environments.
B-2
Heat Stress in Beans: Reproductive Resilience and Sink Strength as Keys to Yield Stability
Jose Polania, Research Scientist, Alliance Bioversity & CIAT
j.polania@cgiar.org
Co-authors: Jorge Aragon, Cristian Tirado, Hector F. Buendia, Jose Polania, Jennifer Wilker
High night temperatures are a critical constraint for common bean (Phaseolus vulgaris) productivity in tropical environments, where minimum temperatures above 22 °C are increasingly common. To dissect physiological mechanisms of heat tolerance, we evaluated diverse germplasm under field conditions at Espinal, Colombia, as well as under controlled greenhouse environments at CIAT – Palmira, Colombia, both characterized by night temperatures consistently above 22 °C.
The germplasm included a panel of 150 elite and historical lines representing both Mesoamerican and Andean gene pools, along with an interspecific RIL population (P. vulgaris × P. acutifolius). Across both environments, the reproductive phase and grain filling were identified as the most vulnerable stages under heat stress. Susceptible genotypes displayed strong reductions in seed number per pod and poor grain filling, leading to smaller, shriveled seeds and reduced visual quality. These findings underscore the sensitivity of reproductive structures and assimilate partitioning processes under elevated night temperatures.
Physiological characterization further revealed that differences in sink strength play a central role in determining yield stability. Genotypes exhibiting greater “sink strength” or mobilization capacity measured as pod harvest index and grain filling efficiency maintained higher grain quality and yield under heat stress. Importantly, image-based phenotyping methods were successfully deployed to rapidly discriminate seed set and grain filling characteristics across genotypes. This approach enables high-throughput identification of tolerant lines with superior sink strength and offers a practical tool for screening large populations under heat stress conditions.
Together, these results demonstrate the combined value of field and greenhouse evaluations in identifying key physiological determinants of heat tolerance in beans. The integration of elite germplasm evaluation, interspecific populations, and image-based phenotyping is providing new insights into the mechanisms of yield stability under high night temperatures. By identifying lines with stronger sink strength and improved reproductive resilience, this work directly informs breeding strategies aimed at delivering heat-tolerant bean varieties adapted to the challenges of climate change.
B-3
Genome-wide Selection for Drought Tolerance in Mesoamerican Bean Genotypes
Vania Moda-Cirino, Research Scientist, Rural Development Institute of Paraná State – IAPAR-EMATER – IDR-Paraná
vamoci@idr.pr.gov.br
Co-authors: SANTOS, Elizeu David; ARRUDA, Isabella Mendonça; SANTOS NETO, José dos; ALVES, Daniel Soares; IÁCONO, Jéssica Delfini de Paula; GEPTS, Paul; MODA-CIRINO, Vania
Water deficit (WD) is one of the main factors reducing the productivity of common bean (Phaseolus vulgaris L.), and the development of drought-tolerant cultivars represents a strategy for ensuring food security. This study aimed to evaluate agromorphological traits and identify genomic regions through genotyping-by-sequencing (GBS) for a genome-wide association study (GWAS) in a Mesoamerican common bean diversity panel, composed by 146 genotypes (G) under WD conditions. The experiment was carried out under protected cultivation at the Experimental Station of IDR-Paraná, Londrina, Paraná State, Brazil, during the 2022 and 2023 fall-winter cropping seasons. The experiment was arranged in a randomized block design with two replications, in two environments (E): water deficit (WD), and well-watered (WW). The genotypes BAT 477 (drought-tolerant), IPR Sabiá, and BRS Pontal (both drought-sensitive) were used as controls. The agromorphological traits were subjected to individual and joint analysis of variance. For GWAS, genotyping was performed using GBS, obtaining 25,011 single nucleotid polymorphism (SNPs) with minor allele frequence (MAF) >0.05. Linkage disequilibrium was calculated using the kinship matrix for each individual chromosome, and the search for favorable alleles and candidate genes was performed using Gene Ontology. Quantitative trait nucleotids (QTNs) were identified using GLM, FarmCPU, and BLINK models, based on the Phaseolus vulgaris reference genome available in Phytozome v13. All analyses were performed using R® and Genes® software. ANOVA showed significant GxE interaction effects for all traits, and the phenotyping identified greater genetic diversity among the accesses under WD conditions. The genotyping identified 23 QTNs, as well as 462 genes associated with the traits evaluated under WD conditions. Among the identified genes, Phvul.010G019100 and Phvul.010G019600 stand out for their roles in mitigating reactive oxygen species (ROS), maintaining photosynthesis through thermal shock proteins (TSPs), and abscisic acid (ABA) signaling. The genes Phvul.009G137900 and Phvul.010G092300 are involved in the activation of ethylene response transcription factors (ERFs), reducing drought stress effects. Molecular techniques combined with phenotyping enabled the identification of genes regulating metabolic pathways related to drought tolerance. In the future, following the validation of these genes, they will be associated with molecular markers and applied in the selection of drought-tolerant genotypes.
B-4
Evaluating Cowpea Heat Tolerance
John Harling, Graduate Student, University of Florida
jharling1@ufl.edu
Co-authors: E. F. Rios, O. Boukar, T. Close, M. Muñoz-Amatriaín, K. Boote, C. Messina, J. Dubeux
Cowpea (Vigna unguiculata L. Walp) is a dual-purpose crop utilized in many regions around the globe, particularly in sub-Saharan Africa but also in the southeastern United States. Cowpea provides affordable protein in people’s diets, fodder for livestock, and it can contribute to soil health. It is a hardy crop that is well adapted to heat and drought, but rising temperatures caused by climate change are causing yield reductions. Therefore, the objective of this study is to contribute towards the development of more heat tolerant varieties of cowpea by identifying germplasm carrying beneficial alleles. This presentation provides a progress report of results from a preliminary differential heated greenhouse trial for cowpea genotypes to be used as controls in further experiments. Nine different genotypes were evaluated in 4 greenhouses with varying temperatures which consisted of mean hourly average temperatures of 24.9, 25.7, 29.5, and 33.5 °C. The experiment was analyzed as a complete randomized design with each greenhouse acting as an environment with 3 replicates of each genotype within each greenhouse. Overall, the increased temperature resulted in earlier flowering time, and occasionally lower seed number and yield in harvested seed weight. However, there was variation in the response to heat stress between genotypes. For instance, the yield of CB3 did not change between the hotter environment (avg. Temp = 29.5 °C) and the cooler environments (avg. Temps = 24.9 and 25.7) while TVu-9265 performed poorly in the hotter environment (avg. Temp = 29.5 °C). All genotypes failed to produce mature pods in the hottest environment (avg. Temp = 33.5 °C). The variation in heat tolerance across genotypes in these preliminary results may contribute to identifying beneficial germplasm for future heat tolerant cowpea varieties.
B-5
Four New Lima Bean RIL Populations for Exploring Agronomic and Seed Appearance Traits
Emmalea Ernest, Extension Fruit and Vegetable Specialist, Cooperative Extension, University of Delaware
emmalea@udel.edu
Co-authors: Emmalea Ernest, Francis Reith, Ekaterina Hampton, Christine Diepenbrock, Antonia Palkovic, Jaclyn Adaskeveg, Varma Penmetsa, Andrew Farmer, Stephanie English, Paul Gepts
Lima bean (Phaseolus lunatus) is an important part of agricultural systems in certain regions of the United States with potential for increased production as a resilient crop, tolerant of heat and drought. Four lima bean RIL populations have been phenotyped in multiple environments/years for traits related to phenology, agronomic performance, plant architecture, growth habit and seed appearance. The populations were recently characterized using genotyping by sequencing.
DE17221 (PI 256417 x PI 534918) is derived from a cross between a photoperiod sensitive vining genotype and a bush, photoperiod insensitive genotype. This population includes lines with diverse seed size, color, pattern and shape. One parent of the population is highly heat tolerant and the other has multiple disease resistance traits.
The parents of the other three populations are all photoperiod insensitive with determinate growth habits to facilitate phenotyping in the field. Each of the determinate growth habit populations has one green seeded parent with the goal of better understanding the inheritance and expression of this trait which is important for succulent lima beans produced in the Mid-Atlantic region.
DE21101 (PI 534918 x ‘C-elite Select’) contains lines with different seed coat colors, heat tolerance, phenology and plant size.
DE21102 (PI 549509 x ‘C-elite Select’) contains lines with different seed coat colors and patterns, different levels of heat tolerance and erect vs decumbent plant architecture.
DE20201 (DE1306583 x DE0802101A) contains lines with different levels of resistance to southern root-knot nematode (Meloidogyne incognita), different seed coat colors and patterns, and different phenology.
B-6
Identifying Candidate Alleles and Landraces for Adaptation of Domesticated Common Bean (P. vulgaris) to Heat and Drought Using an Environmental Genome-Wide Association Study (envGWAS) Approach
Miguel Correa Abondano, Research Associate, Alliance Bioversity-CIAT
m.correa@cgiar.org
Co-authors: Ospina, Jessica; Franco, Jorge; Sonder, Kai; Wenzl, Peter y Carvajal-Yepes, Monica
The Alliance Bioversity-CIAT safeguards the largest collection of Phaseolus germplasm in the world. The collection’s 38,000 accessions, which include wild relatives and all five domesticated Phaseolus species, are an important genetic-diversity resource for current and future breeding efforts given their broad adaptation to a variety of environments. The use of the collection, however, is limited by the lack of germplasm-evaluation data for most accessions. That is why allele-mining approaches that do not rely on phenotypic trait data and can include thousands of accessions are a potentially powerful approach to identify candidate alleles associated with adaptation to dry and warm climates, two of the most limiting abiotic factors for common bean in the Northern Triangle region of Central America (El Salvador, Guatemala, Honduras). In this study, 10,000 P. vulgaris landraces with coordinate data of collection sites were genotyped with the DArTseq platform to call codominant (SNP) and dominant (SilicoDArT) markers. Latitude and longitude of origin were used to extract climate data. These two datasets were subjected to an environmental Genome-Wide Association Study (GWAS). Preliminary results reveal multiple marker-environment associations specific to the Andean and Mesoamerican genepools common bean. Additionally, the genotypic data was used to conduct a comprehensive analysis of population structure across its entire distribution range, leading to the identification of potentially admixed landraces.
B-7
Agronomic Performance, Nutritional Quality and Cooking Traits of Common Bean Genotypes Grown Under Terminal Drought Conditions in South-Central Chile
Kianyon Tay, Research Scientist, Instituto de Investigaciones Agropecuarias (INIA) and Universidad de Concepción
ktayn@inia.cl
Co-authors: Nelson Zapata, Carlos A. Urrea, Miguel Garriga, Abdelhalim Elazab and María Dolores López-Belchí.
Chile winter precipitation expected to decline by more than 40% by the end of the century, this scenario has led to a prolonged mega-drought, marked by a 20–40% reduction in rainfall, diminished Andean snowmelt, and reduced water availability in rivers and reservoirs, particularly during summer months. This drought stress could significantly impacts the agronomic and nutritional properties of common bean (Phaseolus vulgaris L), affecting agronomic traits, the content of bioactive compounds, and their antioxidant activity. This study evaluated the effect of two water regimes—no drought (ND) and terminal drought stress (DS)—on twenty bean genotypes and commercial cultivars over two growing seasons. Productive and quality traits were assessed, including grain yield (GY), number of grains per pod (NGP), hundred-grain weight (HGW), hydration capacity (HC), cooking time (CT), crude protein content (PC), total phenolic content (TPC), and DPPH-based antioxidant activity. Results showed a significant reduction in GY (-22.3 %) and NGP (-61.3%) under DS, while HC increased for most genotypes (+1,9%). Some genotypes, such as 464, 483, and 478, maintained high GY (> 3600kg ha-1) despite reduced NGP. A positive correlation was observed between seed weight and protein content (r = 0.4) under ND and DS conditions. Under DS, most genotypes exhibited increased phenolic compounds and antioxidant capacity, highlighting their potential for stress resilience. The study highlights the intricate relationship between productivity, grain quality, and drought resistance in common beans. These findings could contribute to improving the genotype selection process and development of common bean cultivars, indicating that selecting genotypes for both high yield and quality under water-limited conditions, as well as sustainability and resilience to climate variability in this important crop.
B-8
Assessing the Role of Pollen Release and Stigma Exsertion in Lima Bean Yield
Ekaterina Hampton, Graduate Student, University of Delaware
ekathamp@udel.edu
Co-authors: Emmalea Ernest, Ashish Reddy, Yin Bao
Understanding the reproductive biology of lima bean (Phaseolus lunatus) is critical for improving yield potential, particularly under high temperature conditions. High nighttime temperatures during the flowering stage can disrupt floral development and reduce yield. One observed effect of heat stress is stigma exsertion, where the stigma extends beyond the anthers. This alteration in floral morphology may influence pollination efficiency by affecting pollen transfer and fertilization. Traits such as pollen release and stigma exsertion may influence pollination success, seed set, and ultimately yield.
The objective of this study is to examine the relationship between pollen counts, stigma exsertion, and yield in lima beans in both normal and heat-stressed environments. A diversity panel of 150 genotypes with two replications was grown under controlled greenhouse conditions in the spring, with average daytime and nighttime temperatures of 28°C (82°F) and 19°C (66°F), respectively. From each genotype, three newly opened flowers were sampled in triplicate. Stigmas were carefully extracted by hand, placed into microcentrifuge tubes containing distilled water and a 2% acetocarmine stain, and then mounted on microscope slides. Three unique images per sample were taken, and pollen counts were determined using a deep learning-based object detection model, YOLOv11-X, achieving a mean average precision (mAP) of 0.924. Training annotations were created with Roboflow. Pollen detection and counting process was automated through a Hugging Face Space which outputs results in a spreadsheet (CSV format) for further analysis. Following pod development, mature pods were harvested and evaluated for seed count and dry weight. The same genotype panel was evaluated in the summer under heat-stressed conditions, where average temperatures reached 33°C (91°F) during the day and 27°C (80°F) at night, with nighttime heating used to maintain elevated temperatures.
This ongoing research will provide insight into how reproductive traits such as pollen release and stigma exsertion interact with environmental stress to influence yield in lima bean. Furthermore, the pollen staining, imaging, and automated counting methods developed here may serve as useful tools for future studies on pollination biology and for breeding programs targeting heat resilience.
B-9
Presence of a New Bean Anthracnose Race in Northern Spain
Elisabeth Portilla Benavides, Graduate Student, Regional Service for Agrofood Research and Development (SERIDA)
anaelisabeth.portillabenavides@asturias.org
Co-authors: A. Campa, M. Suárez-Fernández, J.J. Ferreira
Severe anthracnose attacks, caused by the fungus Colletotichum lindumuthianum (Sacc&Magnus), have been observed in northern Spain in 2024, affecting the Fabada market class. These attacks were even observed in the A2806 breeding line, in which the Co-2 resistance cluster was introgressed from Cornell49-242. The most frequent pathogenic variant in this region was the race 38, for which the Co-2 from Cornell49-242 conferred total resistance. Due to the suspected presence of new pathogenic variants, in this study several C. lindemuthianum isolates were collected in 2024 from infected fields and classified into races. The virulence of one of these isolates was investigated by assessing the resistance response of a diversity panel under controlled conditions.
Isolations were made from the seeds or pods of naturally infected bean plants (market class Fabada). Monosporic cultures were obtained from each isolate and characterized using the standard set of 12 differential cultivars and a set of seven Fabada breeding lines. In total, three monosporic isolates (CL232, CL233, and CL234) were selected and classified as races 5 (CL232 and CL234) and race 37 (CL233). Isolates classified as race5 produces intermediate symptoms in Fabada lines carrying the Co-2 cluster: A2806, A4804, X2776, and X4562. A diversity bean panel constituted in the INCREASE project (https://www.pulsesincrease.eu/) was evaluated against race 5. This panel contains 431 lines that represents a wide range of genetic diversity from America and Europe. The results revealed only 71 resistant lines (16,5%), 7 with intermediate reactions (1.6%) and 353 susceptible lines (82 %). This represents higher virulence compared to race 38 which caused susceptibility in only 59% of the INCREASE diversity panel.
This is the first report describing the presence of races 5 and 37 in northern Spain, and due to the virulence of race 5, an anthracnose breeding program should be defined to protect the local cultivar Fabada.
B-10
Population Dynamics of Megalurothrips usitatus in Common Bean in Guatemala
Jose Figueroa-Cerna, Research Scientist, North Dakota State University
jose.figueroacerna@ndsu.edu
Co-authors: Inelda Z. Tuj; Astrid J. Racancoj; Angela N. Miranda
Since 2020, the twice-banded bean thrips (Megalurothrips usitatus Bagnall), a species with a preference for legumes, has been reported in several countries across the Americas. Its presence in Guatemala was confirmed as early as 2021 in the locations of Zacapa, Petén, Chimaltenango, and Baja Verapaz, where it was found damaging both dry and snap beans (Phaseolus vulgaris L.) crops. This study aimed to characterize the behavior of M. usitatus and its impact on common beans. An observational plot of 320 square meters was established with the cultivar ICTA Super Chiva (a Bush-Type variety) in the central highlands of Guatemala (1,800 meters of altitude) during the second semester of 2023 (August to December). A descriptive and quantitative study was conducted, with weekly counts of adult and nymphal populations of M. usitatus. During the reproductive phase of the common bean plants, 30 flowers and pods were evaluated. In addition, a pod damage scale was developed (1 = healthy pod, 9 = severely damaged pod), and grain quality was assessed. Colonization by M. usitatus began 10 days after planting (DAP), with nymphal populations peaking at 31 and 66 DAP. At flowering, the highest density of adults was recorded at 74 DAP, with 27 individuals per flower. Nymphal feeding caused pod deformities and scarring of the pericarp. Pod damage incidence reached 60 %, of which 45 % corresponded to severe damage. Infestation by M. usitatus resulted in a 28 % reduction in seed yield and a decline in commercial grain quality. Increasing pod damage was associated with reduced seed weight and shorter cooking times, leading to uneven cooking. These findings indicate that M. usitatus is a significant threat to common bean production, as it damages multiple plant structures and directly reduces seed yield potential and seed quality.
B-11
Identifying Genomic and Phenotypic Variation amongst Diverse Sclerotinia sclerotiorum Isolates: A Steppingstone towards Improved White Mold Tolerance Screening
Marysia Zaleski-Cox, Graduate Student, Michigan State University
zalesk34@msu.edu
Co-authors: Tanya Copley, Mark Derbyshire, Lone Buchwaldt, Syama Chatterton and Valerio Hoyos-Villegas
White mold, caused by the fungal pathogen Sclerotinia sclerotiorum (Lib.) de Bary (Ss), is a devastating pathogen capable of infecting hundreds of plant species and greatly reducing yields of crops such as common bean (Phaseolus vulgaris L.). Ss isolates vary in their aggressivity thus inoculation of breeding material with multiple, diverse isolates is important to accurately identify white mold tolerance. Our main objective is to understand genomic and phenotypic diversity of Ss isolates collected from geographically distant locations using whole genome sequence data. Seedling inoculations can then be used to identify a comprehensive panel of Ss isolates for white mold screening. To achieve this objective, short read whole genome sequencing of 63 isolates from across Canada was done. Following alignment, polymorphisms were identified and a distance matrix was constructed to build a phylogenetic tree. The four identified subpopulations were verified using Bayesian clustering. One cluster of highly related samples was identified, supporting the need for distantly collected Ss samples in white mold screening. Hundreds of high or moderate impact mutations were identified including within regions encoding necrosis inducing effector proteins such as SsNE3 (SS1G_00872), but their effects on aggressivity require validation. Approximately half of the sequenced isolates have been used to inoculate three common bean lines (G122, USPT-WM-12 and Ex2141-P) representative of the three distinct genetic sources of tolerance to Ss using seedling straw tests. A significant effect of Ss isolate x genetic pool of common bean tolerance has been observed from the resulting data. Further aggressivity data from seedling straw tests are needed before we can recommend a standardized white mold screening isolate panel. In the next stages of this project, we will conduct GWAS on both common bean and Ss to better understand the bean resistance loci required to combat each subpopulation of Ss. Loci significantly associated with tolerance against each Ss subpopulation will be used to create a molecular differential system that may identify tolerance within breeding material.
B-12
Assessing Powdery Mildew Infection in the Tepary Diversity Panel in Puerto Rico
Ihann Rosado, Graduate Student, University of Puerto Rico. Department of Agro-Environmental Sciences, Mayagüez Campus
Ihann.rosado@upr.edu
Co-authors: Consuelo Estevez de Jensen and Timothy G. Porch
Powdery mildew, caused by Erysiphe vignae affects tepary beans (Phaseolus acutifolious) far more than its sister species, common bean (Phaseolus vulgaris). In Puerto Rico, the disease occurs when moderate temperatures and high humidity conditions are prevalent. In Mayagüez and Juana Diaz, PR, disease severity was evaluated in a subset of 30 different cultivated and wild tepary accessions and common beans (controls) from the Tepary Diversity Panel. These were planted under screenhouse conditions and humidity was augmented utilizing humidifiers. The humidity and temperature conditions during the experiment were: 78.9%, 27.3°C in Mayagüez and 76.5%, 27.7°C in Juana Diaz. Disease severity was evaluated at 7, 14 days and 21-day intervals after disease onset, on a 1-9 scale. The common beans ‘Beniquez’ (small white bean) and PR0443-151 (black bean) showed no symptoms or signs of the disease. Tepary beans, TDP-96 (G40117A) and TDP-192 (G40213), TDP-203 (G40223), TDP-234 (G40253) and TDP-344 (PI 440811) showed resistance (DS 1, 1, 1.7, 2.3, 2.3 respectively). Susceptible accessions were TDP-62 (G40080), TDP-264 (G40288), TDP-44 (G40056), TDP-388 (G40121) and TDP-263 (G40287) with a disease severity average of 6.8, 7.3, 7.3, 7.3 and 7.5, respectively. These findings have identified resistant tepary bean accessions that can serve as valuable sources of powdery mildew resistance for tepary and common bean improvement.
B-13
Prediction of White Mold Disease Severity and Symptoms in Dry Beans from UAS Data
John Hawkins, Graduate Student, Michigan State University
hawki345@msu.edu
Co-authors: John Hawkins and Valerio Hoyos-Villegas
White mold is the major yield-limiting disease in dry beans in the US and white mold resistance is a major selection target in many dry bean breeding programs. The typical method of phenotyping white mold severity is to walk the field at maturity and rate disease visually on a 9-point scale based on the proportion of bleached stems. However, these ratings are neither fully objective nor replicable. A single visual measurement at plant maturity also fails to capture white mold development over time. In this research machine learning approaches are being explored to develop UAS imaging-based models for white mold phenotyping which aim to be objective, replicable, and able to incorporate data across multiple time points. The first approach uses a random forest machine-learning algorithm to create an optimal linear model from a large set of vegetation indices collected at different timepoints in order to estimate white mold severity score. A preliminary version of this model, trained on 2024 imaging data, has an RMSE on unseen data of 1.79 and explains 30.64% of the variation. Models were also trained for yield under white mold pressure. The 2024 yield model explained 55.42% of the variation with an RMSE of 6.05 hundredweight per acre on unseen data. Integration of multiple years of data for model training is expected to improve model performance. Other directions for future investigation include the integration of predictors based on plant area and the use of deep learning classifier algorithms to predict the white mold index of plots directly from RGB images.
B-14
Identifying Early Root Traits for Improved Fusarium Root Rot Resistance in Field Grown Phaseolus vulgaris
Asia Hawkins, Graduate Student, Michigan State University
hawki283@msu.edu
Co-authors: Weijia Wang, Paige Smith, Miranda Haus
Fusarium root rot (FRR) is a leading cause of dry bean yield loss in Michigan and across the United States for many crop species. There is little resistance to FRR in dry bean, and what resistance has been identified is quantitative. Our data suggest that early root development in bean seedlings directly impacts disease severity in a controlled environment. Specifically, the length of the apical unbranched zone (LAUZ) of basal roots was shorter in more resistant lines, while susceptible lines showed reduced lateral root formation and increased spacing between the lateral roots and the basal roots. To determine if these findings translate to a field setting, six recombinant inbred lines of P. vulgaris were grown under typical field conditions, with paired rows inoculated with F. brasiliense or with a control. Seedling roots were collected at four, six, eight, and ten days post-planting. The roots were washed, photographed, and certain traits, such as basal root length, diameter, and LAUZ, were measured using ImageJ. Images were then analyzed using RhizoVision Explorer to quantify root architectural traits at each time point. This approach will allow us to identify specific traits associated with developmental timing and increased resistance. Our early analysis shows that variance across the time frame increased as seedlings developed, with days four and six showing tightly clustered root traits, while days eight and ten exhibited greater divergence, particularly along PC1, suggesting that root architecture becomes more variable at later developmental stages. Overall, this dataset helped reveal which early root features may contribute to FRR resistance in field conditions.
B-15
Resistance Gene Mapping and Transcriptomic Profiling of Common Bean in Response to the Infection by Colletotrichum lindemuthianum
Ruifeng He, Research Scientist, USDA-ARS Soybean Genomics & Improvement Lab (SGIL)
ruifeng.he@usda.gov
Co-authors: Larissa F. S. Xavier, Giseli Valentini, Marcial A. Pastor-Corrales, Maria Celeste Gonçalves-Vidigal, Qijian Song
Beija Flor (BF) is an Andean common bean (Phaseolus vulgaris) cultivar with broad-spectrum resistance to Colletotrichum lindemuthianum (CL), the anthracnose (ANT) pathogen of common bean. Fine mapping of the Co-Bf ANT resistance gene in BF was achieved using F3 plants derived from the BF × Crioulo 159 (CR) cross which were genotyped with Kompetitive Allele Specific PCR (KASP) markers. The Co-Bf gene was fine mapped to a 30.2 Kb genomic region on chromosome Pv04, between KASP markers SS333 (342,926 bp) and SS509 (373,114 bp). Three candidate genes were identified within this region, including the Phvul.004G005600 gene that contains NB-ARC and LRR domains associated with plant disease resistance. To investigate the gene expression profiles associated with the ANT resistance, transcriptomic RNA-seq was performed on the BF and CR cultivars inoculated with race 453 of CL. Leaf samples were collected at 24, 48, and 72 hours after inoculation (hai). Total RNA was extracted, and cDNA libraries were constructed and sequenced on the Illumina NovaSeq X Plus platform (150 bp paired-end reads; >6 G raw data/sample). Sequence reads were trimmed and aligned to the G19833 common bean reference genome, using HISAT2. Differentially expressed genes (DEGs) were identified with a >2-fold change and p-value <0.05. Comparative transcriptomics between BF and CR revealed 3309, 3433, and 3681 DEGs at 24, 48, and 72 hai, respectively. We then filtered the DEGs that encode proteins with domains related to pathogen resistance (LRR, NB-ARC, TIR, and Kinase) and a total of 354, 372, and 301 DEGs were identified at 24, 48, and 72 hai, respectively. The Phvul.004G005600 gene, that was previously fine mapped, was expressed at 72 hai. GO, KEGG and transcription factors for each DEG were analyzed. The RNA-seq results revealed multiple DEGs working in response to the ANT pathogen inoculation. These genes were expressed across all 11 chromosomes. The RNA-seq data provides insights into the ANT resistance mechanism. The molecular markers linked to the Co-Bf gene could be used for marker-assisted selection to combine the Co-Bf with other resistance genes in a single cultivar that might confer broad and durable resistance to CL in common bean.
B-16
Diseases Resistance, Adaptability, and Stability of Common Bean Lines in Multi-Environment Trials in Paraná, Brazil
Pedro Soares Vidigal Filho, Research Scientist, Universidade Estadual de Maringá
vidigalfilhop@gmail.com
Co-authors: Jaqueline B. Silva, Maria Celeste Gonçalves-Vidigal, Mariana Vaz-Bisneta, Giselly F. Lacanallo, Andressa G. V. Rosenberg
In Brazil, one of the largest producers and consumers of common bean (Phaseolus vulgaris L.) worldwide, challenges such as diseases, pests, and agro-climatic factors highlight the importance of breeding programs aimed at developing more productive and resistant cultivars. This study evaluated the performance of common bean lines across different environments in the state of Paraná, with the goal of identifying genotypes with stability, adaptability, and disease resistance. Six common bean lines, along with three control cultivars, were evaluated for resistance to anthracnose (Colletotrichum lindemuthianum), angular leaf spot (Pseudocercospora griseola), common bacterial blight (Xanthomonas axonopodis pv. phaseoli), and Curtobacterium wilt (Curtobacterium flaccumfaciens). The experiments were conducted during the 2020, 2021, and 2022 agricultural years in six municipalities of Paraná, Brazil: Guarapuava, Ponta Grossa, Santa Tereza do Oeste, Pato Branco, Maringá, and Campo Mourão (the latter in 2021 only). Adaptability and stability across environments were evaluated using the methodologies of Eberhart and Russell (1966), Lin and Binns (1988) as modified by Carneiro (1998), Annicchiarico (1992), and WAASB (Olivoto et al., 2019), allowing a comprehensive analysis of genotype-by-environment (G×E) interaction. The line LP 09-180 demonstrated consistent tolerance to anthracnose across locations and years, particularly under high disease incidence. Angular leaf spot showed low incidence in the evaluated lines, indicating good tolerance under field conditions. The lines CHC 04-233-2, CHP 12-355-02, LEC 03-16, and LEP 01-16 exhibited tolerance to common bacterial blight in multiple environments, reinforcing their adaptability and resilience under conditions favorable to pathogen development. In contrast, the lines LP 08-186 and LP 09-180 displayed variable responses to bacterial blight, ranging from moderately tolerant to susceptible depending on the environment, which may be attributed to G×E interactions. The results regarding Curtobacterium wilt also indicated differential responses across environments and years. Lines such as CHC 04-233-2 and LEC 03-16 showed superior performance, displaying tolerance in several regions and suggesting enhanced resistance under conditions favorable to the pathogen. Overall, these findings underscore the importance of selecting lines that combine agronomic efficiency with disease resistance, thereby ensuring competitive and productive cultivars adapted to diverse cropping conditions.
B-17
Fine Mapping of the Anthracnose Resistance Gene Co-PA in the Andean Common Bean Cultivar Paloma
Maria Celeste Gonçalves-Vidigal, Research Scientist, Universidade estadual de Maringá
mcgvidigal@uem.br
Co-authors: João V. S. Alves, Maria Celeste Gonçalves-Vidigal, Giselly F. Lacanallo, Mariana Vaz-Bisneta, Larissa F. S. Xavier, Ruifeng H, Qijian Song
Anthracnose, caused by Colletotrichum lindemuthianum, is one of the most significant diseases affecting common bean crops. The pathogen’s high genetic variability poses a major challenge to anthracnose control, making the identification of new sources of broad-spectrum resistance crucial for common bean breeding programs. In this context, the Andean cultivar Paloma is notable for its differentiated resistance to C. lindemuthianum, exhibiting resistance to Andean races 23, 31, and 55, as well as Mesoamerican races 65, 73, 1545, 2047, and 3481. The resistance gene (Co-Pa) in Paloma was previously mapped to chromosome Pv01, between markers SS82 and SS83 at 50,155,987 bp and 50,546,985 bp, spanning a 390.9 Kb region. In the present study, we fine-mapped this region by identifying molecular markers linked to the gene and candidate genes in the reference genome. The experiments were conducted using 100 F2:3 families from the Paloma × AB136 cross to evaluate resistance to anthracnose (race 3481 of C. lindemuthianum). SNP markers from the Illumina BeadChip BARCBean6K_3 were used for genotyping, and KASP markers were applied for fine mapping. A genetic map was constructed with JoinMap 4.0 using the Kosambi function and a LOD score ≥3.0. Candidate genes within the mapped region were identified using the Phaseolus vulgaris v1.0 genome. The results revealed a segregation ratio of 1RR:2RS:1SS (χ² = 0.743; p = 0.68), indicating that resistance to C. lindemuthianum race 3481 in Paloma is controlled by a dominant gene. Fine mapping of 100 F2 plants narrowed the resistance region to 54 Kb, between markers SS97 and SS90. This refined region places Co-Pa in a distinct location from previously identified resistance genes on chromosome Pv01, such as such as Co-1, Co-1⁴, Co-x, Co-AC, Co-1x, and CoPv01ᶜᴰᴿᴷ / PhgPv01ᶜᴰᴿᴷ. Among the six genes identified in the 54 Kb region, Phvul.001G245300, which encodes a protein tyrosine kinase with an LRR domain, stands out as the primary candidate for Co-Pa due to its established role in plant defense responses. These findings provide valuable insights for breeding programs aimed at developing common bean cultivars with enhanced resistance to anthracnose.
B-18
A Domestication Change at PvMYB26 in Common Bean Sheds Light on the Origins of Middle American Agriculture
Burcu Celebioglu, Postdoc, University of California, Davis
bcelebioglu@ucdavis.edu
Co-authors: Jayanta Roy, Andrew Farmer, Stephanie English, Xingyao Yu, Xiaosa Xu, Phillip E. McClean, Paul Gepts, Travis A. Parker
Domestication imposed radically different selection pressures on plants, eventually transforming them into the cultivated forms that support global populations today. Here, we investigate the loss of seed dispersal via pod shattering during common bean (Phaseolus vulgaris L.) domestication. We identified an 8 kb deletion eliminating the transcription start site and promoter of the candidate gene PvMYB26. Mutants express PvMYB26 at < 1% of the level of wild types and produce 44% less pod lignin. RNA in situ hybridization and fluorescence microscopy shows that PvMYB26 is expressed in the lignified fiber layer of pods, while mutants show no visible expression and have a greatly reduced fiber layer. Whole-genome sequencing of 323 accessions revealed that the mutation is nearly diagnostic for domestication status among Middle American common bean, and was associated with a 125 kb hard selective sweep among Middle American domesticates, indicating the gene’s importance in domestication. We also identified a high-frequency PvMYB26 frameshift/premature stop mutation unique to Andean domesticates. Wild haplotypes most like Middle American domesticates are found in eastern Jalisco, Mexico. Our results suggest that West-Central Mexico was the site of common bean domestication and suggest that this region may have been important in the rise of Middle American agriculture.
B-19
The Pan-GS Consortium: Leveraging Collective Breeding Program Data for Durable Genomic Selection Models
Lovepreet Singh, Postdoc, Michigan State University
singhlo2@msu.edu
Co-authors: Robert McGee, Isabella Chiaravallotti, Jamie Larsen, Mohsen Y. Najafabadi, Peter K. Pauls, Lyndsay Schram, Karen Cichy, Phil N. Miklas, Anfu Hou, Parthiba Balasubramanian, Evan M. Wright, Juan M. Osorno, Jose F.-Cerna, Diego Jarquin, Julia’ G-A. Velasco, Virginia M. Moore, Erika F. Everest, Carlos Urrea, Christine Diepenbrock, Bodo Raatz, Timothy Porch, Jennifer Wilker, Winnyfred Amongi, Clare Mukankusi, Valerio Hoyos-Villegas
Breeding for yield in common beans (Phaseolus vulgaris L.) is challenging because it is a complex quantitative trait. Genomic selection (GS) can accelerate genetic gain of the complex quantitative traits by capturing minor allelic effects using genome-wide markers. When paired with phenotypic data from a training population, a trained predictive model can predict genomic-estimated breeding values (GEBVs) within a breeding population. Low GS predictive accuracies are however common with small training populations, with low genetic diversity, and low genetic relatedness to the breeding population. To overcome these limitations, the goal of this project is to combine data across multiple bean breeding programs to train a ‘Pan-GS’ predictive model that should be more robust, accurate, longer-lasting, and enable greater genetic gain than single breeding program-trained predictive models. A meta-phenotypic dataset from 15 common bean breeding programs was assembled, totalling 3,811 breeding lines, 232 checks, one to two years of pre-registration field data across 36 locations in Canada, USA, Puerto Rico, Colombia, France, Spain, and Uganda, within which we will report on some interesting GxE interactions. To reduce genetic redundancy with this training population, the highest yielding sibling from each family was identified based on the highest BLUP value. Using this approach, a final training population of 1419 lines (1,228 breeding lines and 191 checks) was curated, balanced across the Meso-American (n = 911) and Andean (n = 508) gene pools, from all major market classes to minimize class-specific bias in model training, and some rarer market classes. Genomic DNA from the training population; genotyped with the BARCBean12K SNP array, phenotype, and pedigree data are currently being used to train and compare the prediction accuracies of models trained on the complete training population, by gene pool, by closely related market-class, and by single market-classes. Preliminary data on model training optimization and prediction accuracies will be presented. Using this meta-dataset, we will also be able to identify shared and rare alleles between breeding programs. The end goal is to develop the best performing Pan-GS predictive model that, by being publicly available, will remove the major barriers for bean breeders worldwide to perform GS.
B-20
Genetic Variability and Relationships of Yield and Yield Related Traits of Common Bean (Phaseolus vulgaris L.) Genotypes Locally Grown and in Burundi
Eric Nduwarugira, Graduate Student, Institut des Sciences Agronomiques du Burundi (ISABU)
nduweric2003@gmail.com
Co-authors: Susan Nchimbi-Msolla, Paul M. Kusolwa, Teshale Assefa and Clare Mugisha Mukankusi
Common bean (Phaseolus vulgaris L.) is a vital food legume contributing to nutrition, food security, and income. Assessing genetic variability and trait associations is essential for breeding, yet limited information on locally grown genotypes in Burundi constrains their integration into national programs. This study evaluated genetic variability and relationships among yield-related traits in 83 bush and 84 climbing bean genotypes. Field trials were conducted at ISABU’s Bukemba, Karusi, and Murongwe stations using an unbalanced lattice design with three replications during the 2024A and 2024B seasons. Analysis of variance revealed highly significant differences (P < 0.001) among genotypes for all traits, indicating considerable diversity and strong genotype-by-environment interactions. Genetic parameter estimates showed high heritability and genetic advance for hundred seed weight (HSW) in climbing beans (h² = 56%, GA = 9.95 g, GAM = 31%), suggesting suitability for direct selection. In bush beans, plant height (PH_cm) (h² = 64%, GA = 43.51 cm, GAM = 62%) and HSW (h² = 52%, GA = 9.91 g, GAM = 33%) were identified as key selection traits. By contrast, grain yield (GYD_kgha⁻¹) and days to maturity (DTM) exhibited low heritability (<35%), reflecting strong environmental influence. Both bush and climbing beans showed positive associations between yield and PH_cm, number of pods per plant (NPP), number of seeds per pod (NSP), pod harvest index (PHI), and harvest index (HI). Climbing beans exhibited stronger genotypic correlations but greater sensitivity to environmental stress, whereas bush beans showed more stability. Path analysis indicated that NPP (0.31, 0.22), NSP (0.20, 0.23), PH_cm (0.25, 0.10), and HSW (0.22, 0.20) had the strongest direct positive effects on yield in bush and climbing beans, respectively. Given the low heritability of grain yield (<30%), indirect selection was recommended. In bush beans, promising indirect paths included NPP via NSP (0.11) and NSP via PH_cm (0.12), while in climbers they included NSP via PH_cm (0.06), HSW via pod length (0.08), NPP via DTM (0.06), and pod length via HSW (0.07). These findings highlight high-heritability traits and yield components as priority targets for breeding programs to develop climate-resilient, high-yielding bean cultivars adapted to Burundi’s diverse agro-ecological zones.
B-21
Exploring the Genetic Control of Seed Coat Color in a RIL Population Derived from a Red x Red Cross
Elisabeth Portilla Benavides, Graduate Student, Regional Service for Agrofood Research and Development (SERIDA)
anaelisabeth.portillabenavides@asturias.org
Co-authors: C. García-Fernández, A. Campa y JJ. Ferreira
Seed coat color in common bean (Phaseolus vulgaris L.) results from the accumulation of pigments, mainly phenolic compounds, and is an important consumer-related trait associated with human health benefits and response to biotic stresses. Classical inheritance studies have described a complex network of major genes that control color (complex C and genes P, C, V, R, Prp, G, and B, among others) and distribution (complex C, Z, T, Bip, and Fib). A RIL population derived from the two red-seeded genotypes AB136 and MDRK (population ABM) was developed and segregation for seed color was observed. The objective of this study was to investigate the genetic control of red seed color in this population.
A linkage genetic map containing 796 SNP distributed across the eleven bean chromosomes was developed in the ABM population. Qualitative evaluation of seed coat color revealed three phenotypic classes: red parental phenotype (N=52); dark purple (N=30); and pink (N=27). The observed segregation fitted a 2:1:1 ratio (χ2=0.39, p=0.82), suggesting the involvement of two independent genes, where the presence of the parental genotype at both loci results in dark purple color, and the absence at both loci leads to a pink color. Contingency tests were used to locate the genetic regions significantly associated with seed color (p<0.001). Considering the two phenotypic classes black and no black, two regions significantly associated were detected on chromosomes Pv02 (48.21-49.64 Mb) and Pv08 (0.73-4.97 Mb). Considering the two phenotypic classes pink and non-pink, two regions were detected in Pv02 (32.79-43.10 Mb) and Pv08 (1.82-3.13 Mb). Interestingly, when the classes parental (red) and non-parental (non red) are considered, only a region on Pv02 (37.42-39.90 Mb) is detected.
These results suggest the involvement of two major seed color loci in the ABM population, one is located on Pv08 in the genetic position of the C complex, and the other is located at the end of the chromosome Pv02, where the B gene was mapped. Moreover, this study allowed us to understand that for the production of phenolic compounds, two different metabolic pathways can lead to the same seed coat color phenotype.
B-22
Identification of Anthroacnose Resistance Loci Introgressed from P. coccineus into P. vulgaris Combining BSA and WGS
Elisabeth Portilla Benavides, Graduate Student, Regional Service for Agrofood Research and Development (SERIDA)
anaelisabeth.portillabenavides@asturias.org
Co-authors: A Gaiti, A Campa, JJ Ferreira
Bean anthracnose is a race-specific disease, in which numerous pathogenic variants have been described. Phaseolus coccineus L. (Pc) exhibits a high level of resistance to this disease and other common diseases that affect Phaseolus vulgaris L. (Pv). Lines (HYB) derived from the interspecific cross Pv × (Pv × Pc) were developed at SERIDA. The goal of this study was to investigate the inheritance of resistance to race 38 identified in the HYB028 line, as well as to identify the genomic regions responsible for this resistance.
A segregating population was developed from the cross HYB28 (resistant to race 38) × Midas (susceptible). The obtained F2 plants were multiplied in greenhouse. The response to race 38 was evaluated under controlled conditions with at least 16 F3 seedlings per F2 plant. The observed F2 segregation was 21 resistant, 40 heterozygous, and 24 susceptible, which fitted the expected ratio for one gene (χ²1:2:1= 0.21, p =0.78). F2:3 families with the heterozygous genotype showed segregation that conformed to a 3 resistant:1 susceptible ratio, indicating the dominant nature of this resistance. Two bulks (Resistant & Susceptible) with 10 F2 plants each were established and genotyped using whole-genome sequencing (WGS). Sequencing reads from different genotypes were aligned using the bean genome G19833 v2 (ncbi accession GCA_000499845.2). A total of 5,766,560 SNPs were detected and reduced to 5,432,040 SNPs after filtering (SNP with missing values or physically close positions were removed). Bulked segregant analysis was implemented using the package QTLseqR, which revealed two statistically significant regions on chromosomes Pv07 and Pv11, likely associated with anthracnose resistance. The variants identified by WGS can be used to develop molecular markers to be tested in the F2 population for their involvement in resistance by linkage analysis.
The innovative aspect of this work is the use of P. coccineus as a donor to broaden resistance to anthracnose and the combination of BSA and WGS as a rapid and cost-effective approach for identifying genomic regions associated with complex traits.
B-23
Developing Specialty Dry Beans for the Northeastern US using Evolutionary Participatory Plant Breeding
Emily Fratz, Graduate Student, Cornell University
ef389@cornell.edu
Co-authors: Jamie Crawford, Virginia Moore
Dry beans are a nutritious crop, rich in protein, fiber, and micronutrients, offering a shelf-stable, plant-based protein source with strong potential for regional food systems in the Northeastern US. In New York, rising consumer demand for heirloom and specialty beans has not been met by regional supply. This project aims to address that gap by improving regionally adapted dry bean varieties, particularly those suited for small-to-medium scale organic production. By doing so, we seek to increase economic resilience for small farms, diversify crop rotations in vegetable operations, and enhance the availability of local plant-based protein. Participatory plant breeding (PPB) methods are being used to develop dry bean populations, a decentralized approach that centers farmers in the breeding process. The majority of farmers involved are women or gender minorities, groups often underrepresented in agricultural research. We are using a modified evolutionary plant breeding strategy, with breeding targets including regional adaptation, disease resistance, climate resilience, suitability for organic systems, and upright bush architecture, with additional goals driven by farmer and researcher priorities. Anticipated outcomes include the development of dry bean populations, strengthened community among growers and researchers, and enhanced farmer confidence and skills in plant breeding and research. In 2024, crosses were made between commercial market classes and specialty and heirloom varieties that performed well in trials at the Cornell Organic Research Station in Freeville, NY. In June 2025, 74 F2 breeding populations were planted at the Cornell Organic Research Station. A subset of the breeding populations were planted on six farms in diverse climates across the Northeast (Finger Lakes, Hudson Valley, and Long Island, NY, Central Vermont, Western Massachusetts, Western Pennsylvania). This season, data on stand count, flowering time, growth habit, disease and pest pressure, pod height, lodging and yield will inform decisions on populations to advance. Selection will continue for several growing seasons, both at the research station and with farmer-breeders, before moving material to the evaluation stage. All intellectual property decisions will involve the participating farmers.
B-24
Understanding Heat Tolerance in Phaseolus vulgaris Through Comparative Genomics
Leah Tomey, Graduate Student, University of Georgia
leah.tomey@uga.edu
Co-authors: John P. Hamilton, Kathrine Mailloux, Joshua C. Wood, Brieanne Vaillancourt, Tim Porch, C. Robin Buell
Common bean (Phaseolus vulgaris L.) is the third most important legume species being highly valued as a plant protein source for human consumption. Common bean yield is heavily impacted by biotic and abiotic stressors including heat stress. However, a limited number of cultivars are reported as being tolerant to heat. To understand the genetic mechanisms underlying heat tolerance in common bean, we generated de novo genome assemblies of multiple heat sensitive and heat tolerant cultivars of P. vulgaris. Oxford Nanopore Technologies long reads were assembled using hifiasm yielding highly contiguous genome assemblies with total assembly sizes ranging from 567Mb to 596Mb. Assemblies were then scaffolded into 11 chromosomes using RagTag with the common bean GI 19833 reference genome. All genome assemblies are highly contiguous with N50 scaffold values of at least 50Mb and show high completeness with least 99% complete Benchmarking Single Copy Orthologs. GENESPACE, a comparative genome analysis tool was used to visualize, compare, and identify orthologs and synetlogs between the genome assemblies. The data generated from this study will be useful for identifying genes and genomic regions of interest for heat stress related traits in common bean to be used to improve breeding programs.
B-25
Genomic Prediction of Agronomic, Sensing-based, and Grain Macronutrient Traits in Cowpea Across Diverse Environments
Sassoum Lo, Research Scientist, University of California, Davis
ssslo@ucdavis.edu
Co-authors: Jonathan M. Berlingeri, Hamid Kamangir, Margaret Riggs, Astrid Lao, Antonia Palkovic, Alfred Ozimati, Isaac Onziga Dramadri, Patrick Obia Ongom, Ousmane Boukar, Bao-Lam Huynh, Brian Bailey, J. Mason Earles, Christine Diepenbrock
Cowpea (Vigna unguiculata [L.] Walp.) is a major grain and leafy vegetable legume crop in sub-Saharan Africa and other regions of the world. To develop varieties with improved yield, nutritional quality, and environmental adaptability under shifting climatic conditions, this study aims to evaluate the performance of genomic prediction (GP) models across diverse target environments.
To achieve this, a cowpea multi-parent advanced generation intercross (MAGIC) population, comprosing 305 lines, was evaluated across an extensive environmental gradient in the United States (northern to southern California) and Africa (Nigeria and Uganda). Due to photoperiod sensitivity, only 168 early-flowering lines were evaluated under the long-day conditions of California, whereas the full set of 305 lines was assessed in short-day and equatorial environments. Several agronomic traits, including flowering time, grain yield, and hundred-seed weight, as well as sensing-based traits (leaf detection, flower detection, pod detection, plant area, and plant height), have been assessed. Additionally, seed nutritional composition was assayed using custom near-infrared spectroscopy calibrations. The predictive ability of several GP models (GBLUP in TASSEL, RR-BLUP in R, and CHiDO) was compared across traits and environments.
Preliminary results showed moderate to high prediction accuracies, with notable variation across models, traits, and environments. Flowering time displayed model-specific differences, with prediction accuracy exceeding 0.7 in some environments. Protein content showed moderate prediction accuracy (0.3-0.6) across locations. This study will help gain new insights into the potential of GP to accelerate selection for both agronomic performance and nutritional quality in cowpea across diverse environmental conditions within a target population of environments (TPEs), and to leverage datasets as collaborative teams across TPEs.
This work, conducted through international collaboration by a diverse team, supports efforts to accelerate the development of climate-resilient and nutritionally dense grain legume varieties in the US and target countries in Africa.
B-26
On Farm Genomic Selection in Common Bean (Phaseolus vulgaris L)
Teshale Assefa Mamo, Research Scientist, CIAT
teshale.mamo@cgiar.org
Co-authors: Christian R. Werner, D. G. Dieguez, G. Atlin, G. Nyakunga, M. Benedict, K. Kisaka, R.H. Abdilah, H. Chambea, E. Nchanji, N. Kuboja, J. Kwesiga
Beans are critical food security and income-generating crop for over 400 million African farmers and poor urban consumers. Beans are key source of protein, carbohydrate, and micronutrients in Sub-Saharan Africa. Management of beans by African farmers is constrained by the inability to afford yield-maximizing and protecting inputs. Approximately half of African farmers apply no fertilizer at all. Weed pressure is usually higher on the farm than on the research station. The objective was estimating genetic correlations between on farm and on station trials and population development. 426 Stage-1 breeding lines evaluated on 309 farmers field across the bean-growing regions in Tanzania, including the Southern, Northern, and Lake regions, with each region containing 103 on farm sites. Three plots per farm site were used as incomplete blocks, to ensure robust data collection under real farmers conditions. Results indicate the Lake region (LR) and Southern region (SR) share similar target population environments. Negative genetic correlation found between Northern region (NR) and SR and LR, suggesting the necessity of separate breeding pipeline for NR.
Results also show better performance on-station than on-farm. The mean performance across on-station trials was 1.70 tons/ha, with Babati and UYOLE research stations being the highest and lowest performing, respectively. The average performance across the on-farm regions was 0.97 tons/ha, with the SR and NR being the highest and lowest performing on-farm trials. The genomic additive genetic variance was higher on-station trials compared to on-farm trials. The heritability was higher on-station (ranging from 0.85 to 0.92) compared to on-farm, the heritability on-farm were still relatively high (0.67 to 0.87). This study underscores the importance of tailoring breeding pipelines to the unique genetic and environmental conditions of target regions, leveraging both on-station and on-farm data to optimize genetic gains and improve bean yields for smallholder farmers.
B-27
Toward a Navy Bean Pangenome for Dissecting Common Bacterial Blight Resistance and Tepary Bean Introgressions
Mohammad Erfatpour, Research Associate, University of Guelph
erfatpom@uoguelph.ca
Co-authors: Gregory Perry, Maryam Vazin, Karl Peter Pauls
Common bacterial blight (CBB), caused by Xanthomonas axonopodis pv. phaseoli (Xap) and Xanthomonas fuscans subsp. fuscans (Xff), is a major constraint on dry bean production worldwide. The genetic control of CBB resistance in common bean (Phaseolus vulgaris L.) is complex. Although the major QTL on Pv06 (BC420), Pv08 (SU91), and Pv10 (SAP6) have been characterized, none confer complete resistance to aggressive Xap or Xff strains. To address this limitation, some of the first navy bean cultivars with interspecific introgressions from tepary bean (P. acutifolius A. Gray, PI 440795) were developed (e.g., OAC Rex, Apex, Mist, and Rexeter) at the University of Guelph, providing enhanced resistance to CBB. However, the extent of P. acutifolius genome integration in these cultivars and the potential presence of additional resistance- or agronomically relevant alleles remain unclear.
To address this, we are constructing a navy bean pangenome using long-read sequencing of 11 cultivars with contrasting CBB responses. Resistant cultivars carrying P. acutifolius introgressions and susceptible cultivars (e.g., Ex Rico 23, Gryphon, Laser, Cruiser, AC Compass, Envoy, and T9905) are being sequenced on the Oxford Nanopore PromethION 24 platform at 60× coverage. These assemblies will be integrated into a graph-based pangenome to capture the full spectrum of genomic diversity, including structural variants and presence–absence variation.
Our objectives are to (1) validate the previously reported Niemann–Pick–associated locus (now annotated as a patched family protein) for CBB resistance across multiple navy bean backgrounds, (2) quantify the genomic contribution of P. acutifolius to resistant cultivars, and (3) identify novel candidate genes and allelic variation that may contribute not only to disease resistance but also to enhanced resilience to biotic and abiotic stress.
This work will establish the first navy bean pangenome, providing a foundation for genome-wide association, comparative genomics, and marker discovery. Ultimately, this resource will accelerate the development of improved cultivars with durable CBB resistance and inform breeding strategies to balance resistance gains with potential introgression-associated trade-offs.
B-28
Identifying QTL Associated with Popping Trait in Pop Bean
Price Akiina, Graduate Student, University of Wyoming
pakiina@uwyo.edu
Co-authors: Ilyas Ahmad, Donna K. Harris
Pop bean (Phaseolus vulgaris) is a unique Andean bean that expands and produces a pop sound while roasting. It contains high protein, calcium, iron, fiber, and low fat, making it a delicious and nutritious, cholesterol-free snack. The nutritional properties of pop beans make them suitable as a healthy snack.
Unlike popcorn, research is needed to understand the gene(s) controlling the popping trait in pop beans to facilitate pop bean production in the United States.
We developed an F4 mapping population to identify genomic regions controlling the popping trait to achieve our objective.
The parents and 150 lines were subjected to genotype-by-sequencing (GBS) and aligned to the Phaseolus vulgaris UI111 v1.1 genome. The beans were roasted in sand for 90 seconds using an iron cast skillet. Quantitative trait loci (QTL) mapping will be performed to identify regions of the genome that are statistically associated with the popping trait. Further applications and next steps will be discussed.
B-29
Harnessing Genomic Selection and Farmer Participation to Develop Superior Common Bean Varieties for Tanzania
Hector Buendia, Research Scientist, CIAT
h.f.buendia@cgiar.org
Co-authors: Teshale Mamo, Steve Beebe, Jennifer Wilker
The common bean (Phaseolus vulgaris L) is a critical source of protein, essential micronutrients such as iron and zinc, and fiber. Beans are a staple in local diets across Sub-Saharan Africa, and in Tanzania bean consumption is 15.7 kg per capita, among the highest in the world. Bean production in Tanzania is 1,48 million tons a directed at meeting domestic and export needs. In recent years bean breeding programs have faced the challenge of developing demand-led varieties while being more efficient with resources. Genomic selection (GS) is the best methodology for accelerating a breeding cycle. GS has advantages over phenotype-based selection: 1) cost per breeding cycle, 2) time required for variety development.
The accuracy of GS depends on several genetic factors, including the genetic architecture of the trait, population size and diversity, trait heritability, and the presence of many markers that are not in linkage disequilibrium.
For the present study “On farm GS - Beans” population was built at CIAT-Palmira using diverse genetic sources, including elite bush and climbing lines from the Andean and Mesoamerican P. vulgaris gene pools and interspecific lines. Selection of lines for inclusion in the panel was guided by market segment preferences for medium-large seeded red, red-mottle and yellow grain classes.
The GS population consisted of 262 Red and Red mottled lines and 188 yellow lines; all derived from three-way and double crosses. This breeding strategy was used to improve grain yield and recover the commercial-class grain trait.
All lines were evaluated in two environments drought and control (without stress) using a spatial model to account for row and column effect. This methodology minimizes environmental noise allowing for the selection of the lines with high performance under drought stress (more than 1 ton/ha) and grain shape that met commercial standards
The GS population was used in a study to validate the performance of advanced lines. This was done by directly comparing their results under research station conditions with their performance under farmer field conditions. A key component of this research was a participatory model, which incorporated direct farmer input and collaboration.
B-30
Use of Molecular Markers to Fix the bc-3 Gene in Andean Large-seeded Climbing Beans
Hector Buendia, Research Scientist, CIAT
h.f.buendia@cgiar.org
Co-authors: Elizabeth Portilla; Bodo Raatz; Christian Cadena; Jennifer Wilker
The common bean (Phaseolus vulgaris L.) is considered a major source of nutrition, providing protein, fiber, carbohydrates, vitamins, and micronutrients in Latin America and Sub-Saharan Africa. However, diseases and viruses cause significant yield losses (20–100%). The Bean common mosaic virus (BCMV) is one of the most limiting viruses for the Andean gene pool, causing large yield reductions. BCMV is transmitted by infected seed and it is a significant problem for seed production.
The importance of climbing beans is their yield potential, which can be up to three times higher per unit area compared to bush beans, making them an excellent option to improve bean production in growing areas.
In this study using conventional breeding techniques, the breeding strategy was three-way. It was used to improve grain yield and recover the commercial-class grain trait. In early generations (F2) the marker-assisted selection (MAS). This allowed us to identify in early generations the gametes carrying the gene in the absence of stress (BCMV), accelerating the process of generating resistant lines and at a lower cost. was involved in identifying the introgression of the bc-3 gene most accepted climbing bean market classes for Africa. All lines were evaluated in the mid altitude environment using a spatial model to account for row and column effect. This methodology minimizes environmental noise allowing for the selection of the lines with bc-3 gene and yield potential up to 5 tons per hectare, and commercial grain (Mottled Red, Sugar), surpassing yield of the commercial check between 20-40%.
B-31
Expanding Genetic Diversity in Common Bean Through EMS Mutagenesis: A Novel Platform for Variant Discovery and Candidate Genes Validation
Angelo Gaiti, Graduate Student, University of Milan
angelo.gaiti@unimi.it
Co-authors: Chiara Russo, Eleonora Cominelli, Giulio Testone, Paolo Cozzi, Dario Paolo, Massimo Galbiati, Alessia Losa, Tea Sala, Elena Avite, Paolo Andreatta, Carlo Pozzi, Francesca Sparvoli
Common bean (Phaseolus vulgaris L.) is the primary grain legume consumed directly by humans worldwide and is increasingly valued for its high nutritional quality, making it a prime target for genomics research. Despite its importance, genomic resources for this species remain limited.
Within the PhasTILL and AGRITECH projects, an EMS-mutagenized population of 2,345 M1 lines was recently developed from an Italian cultivated variety (Meccearly dwarf Borlotto type, Verisem Italia). This population represents a unique resource for isolating mutants with desirable traits. To date, M3 seeds obtained from approximately 2,300 M2 plants derived from 590 M1 lines are available. Phenotypic screening in M2 and M3 generations has identified major alterations in germination and plant morphology. DNA has been collected from all M2 plants, and whole-genome sequencing (WGS) is ongoing to characterize induced variants.
A pilot WGS study was conducted to optimize variant calling pipelines for EMS-induced mutations. The dataset included two wild-type controls and six M2 pools of four individuals each (4X pooling) sequenced at 40X coverage, and four individual plants from two M2 pools sequenced at 10X coverage. Variant calling was performed using five strategies varying in caller, pooling, and merging approaches, including Freebayes (single-run, batch-wise, octaploid pooling) and HaplotypeCaller with a per-sample gVCF workflow. After filtering, an average of 26,713 variants per sample was identified. HaplotypeCaller detected approximately five times more canonical variants than Freebayes (85,657 vs 11,110–23,363), but with lower confirmation rate in 1X samples (53–62% vs 92–99%). Importantly, 53–94% of canonical variants specific to two M2 pools and confirmed in 1X samples were also recovered by HaplotypeCaller.
These results demonstrate that EMS-induced variants can be reliably identified using short-read WGS of both pools and individual mutants. The observed differences in variant calling highlight the importance of balancing sensitivity and precision. In conclusion, this population broadens genetic diversity in common bean and represents a valuable platform to validate candidate genes, identify potential targets for susceptibility to diseases and accelerate functional genomics studies aimed at improving performance, productivity, and nutritional quality.
B-32
Genomics for Enhanced Anthracnose Resistance in Ontario Dry Beans
Aashvi Patel, Graduate Student, University of Guelph
aashvi@uoguelph.ca
Co-authors: Sajal Ahlawat, Mohsen Yoosefzadeh Najafabadi.
Anthracnose, caused by the fungus Colletotrichum lindemuthianum, is a significant threat to dry bean (Phaseolus vulgaris) production in Ontario and globally, leading to significant yield losses. This study focuses on better understanding the genetic basis of anthracnose resistance in a diverse panel of dry bean genotypes. A total of 195 small seeded dry bean genotypes are being evaluated across two environmentally distinct field sites, using a replicated lattice design over the 2025 and 2026 growing seasons to capture both genetic and environmental variation in resistance responses. Phenotypic assessments include detailed scoring of anthracnose symptoms as well as measurements of key agronomic traits such as plant height, maturity, and harvestability, applying standardized protocols developed by the Ontario Pulse Crop Committee. Preliminary findings indicate a range of disease responses across sites, with most genotypes exhibiting low to moderate infection severity and negligible impact on harvestability or maturity. Notably, no genotypes showed complete infection this may have been due to unideal environmental conditions. There were certain cultivars that showed high resistance to anthracnose in both fields, and some that only showed high resistance in one field. Plots with a severe infection had higher initial plot heights but there was variation in plot height with less severe infection. Subsequent analyses will integrate high-density genotyping data generated via genotyping-by-sequencing to conduct genome-wide association studies (GWAS). These studies will identify single nucleotide polymorphisms (SNPs) significantly associated with anthracnose resistance, providing molecular markers for marker-assisted selection. The identification of genomic regions linked to resistance will enhance the dry bean breeding program’s capacity to develop cultivars with improved and durable anthracnose resistance. This genomics-driven approach will accelerate the development of resilient bean varieties for sustainable production in Ontario and similar environments.
B-33
35 Years of Genetic Gain in Dry Bean Breeding: Evidence from Multi-Environment Trials in North Dakota and Minnesota
Nusrat Khan, Postdoc, North Dakota State University
nusrat.khan@ndus.edu
Co-authors: Sara D. Long, Jose C. Figueroa-Cerna, Juan M. Osorno
Plant breeding programs rely on multi-environment trials (METs) to evaluate genotypes across diverse locations and years and identify lines with superior performance. In addition, long-term MET datasets provide a unique opportunity to quantify genetic progress achieved through breeding over time. In this study, 35 years (1987–2022) of North Dakota State University (NDSU) dry bean (Phaseolus vulgaris L.) breeding trials were analyzed to assess seed yield (SY) trends across five major market classes—pinto, navy, black, kidney, and great northern—grown at nine locations in North Dakota and Minnesota. The analysis aimed to estimate the annual genetic gain in SY for each market class. Best linear unbiased estimates (BLUEs) for SY were calculated for each genotype × environment × year combination, and linear regression on year was used to determine yield gain per year within each class. Breeding trial yield trends were also compared with USDA-NASS dry bean production records to validate the representativeness of the NDSU data in reflecting regional yield progress. Results showed positive genetic gains in SY for most market classes. Kidney beans showed the highest annual yield gain (+13.8 lbs acre⁻¹ yr⁻¹), followed by pinto (+8.7 lbs), black (+5.9 lbs), and navy (+1.1 lbs). Over the 35-year period, dry bean market classes demonstrated consistent and encouraging yield gains. This outcome reflects the long-term effectiveness of selection in improving dry bean productivity. These findings underscore the sustained genetic gains achieved by the breeding program and illustrate the value of retrospective MET analyses for understanding genetic improvement. Notably, the continual yield gains indicate that NDSU’s breeders have successfully developed higher-yielding cultivars without compromising performance. Such long-term insights can inform breeding priorities and guide future breeding strategies to sustain yield gains.
B-34
Genetic Diversity of Andean Common Bean Cultivars and Breeding Lines Revealed by Morphoagronomic Traits
Isabella Arruda, Postdoc, Rural Development Institute of Paraná State – IAPAR-EMATER (IDR-Paraná)
isabellamendonca92@gmail.com
Co-authors: Elizeu David dos Santos; José dos Santos Neto; Vania Moda-Cirino
The common bean (Phaseolus vulgaris L.) is a legume widely cultivated, and constitutes one of the main sources of food for human consumption, with high nutritional value and significant socio-economic importance. Andean beans stand out for their diversity of grain colors, shapes, and sizes. In this context, genetic diversity studies are essential to understand the relationships among accessions adapted to specific edaphoclimatic conditions as well as to guide plant breeding strategies. This study aimed to evaluate the genetic diversity of Andean common bean cultivars and breeding lines through morphological and agronomic descriptors. The experiment was conducted during the 2024/25 rainy season at the Experimental Station of IDR-Paraná, in Londrina, Paraná State, Brazil. A randomized block design with three replications was used with plots consisting of four 4-m rows spaced 0.5 m apart. 13 breeding lines and 7 cultivars (BRS-Embaixador, IPR-Cardeal, BRS-Ártico, IPR-Garça, IAC-2157, BRSFS-311, and IAC-2153) were evaluated. Characterization was carried out using 66 traits, of which 23 were quantitative, and 43 qualitative. Quantitative traits were subjected to analysis of variance (ANOVA, p ≤ 0.05) by Genes software. Subsequently, all traits were subjected to genetic dissimilarity analysis through Ward's distance. ANOVA indicated a significant effect for 18 out of the 23 traits. Three groups were identified in the genetic dissimilarity analysis. The first group consisted of genotypes that demonstrated the highest values for pod length, width, and thickness; number of locules and seeds per pod; total number of seeds; seed length and average yield of 2,990.72 kg ha⁻¹. The second group presented the longest total growth cycles and the highest values for first pod insertion height, number of nodes, and total number of pods per plant. However, it showed the lowest yield with an average of 2,314.47 kg ha⁻¹. Lastly, the third group demonstrated the highest values for plant height, seed thickness and width. Morphoagronomic traits allowed the distinction of common bean cultivars and breeding lines. It highlighted significant genetic diversity among the evaluated genotypes, which are potential parents for breeding programs aimed at developing cultivars with high yield potential and broad adaptation to different edaphoclimatic conditions.
B-35
Evaluating a Breeding Panel of Historical Common Bean (Phaseolus vulgaris) Lines for Effects of Indirect Selection
Grace Sidberry, Graduate Student, Michigan State University
sidberr1@msu.edu
Co-authors: C. Robin Buell, K. A. Cichy, Miranda J. Haus
Historically, Michigan breeding efforts in common bean focused on curating a consistent, upright plant architecture better suited to direct harvesting than the vining and climbing growth habit of wild bean. This scenario is unique because often phenotypic diversity is higher in the wild populations of a crop, whereas in this instance phenotypic diversity for growth habit is higher in the domesticated populations. Breeding efforts have resulted in changes to above ground structural and nutritional traits, but changes to the below ground root architecture due to indirect selection remain unknown. To characterize effects on root systems, 16 historical common bean lines from three market classes, (wild, navy, black, and small reds) were grown in the field during 2020 and 2021. Observed maturity and growth habits were recorded, and plants were sampled for leaf and seed nutrition measurements. Root systems of sampled plants were excavated and scanned to produce images later processed in Root Painter and Rhizo Vision for root architectural traits. Stem diameter has increased overall, reflecting efforts to select in favor of thicker stems, a trait associated with desirable upright growth habits. Plant length has shortened over time because of selection and breeding for less vining growth habits. Exploratory data analysis shows root traits clearly separate by experimental data year compared to leaf and seed nutrition. Principal component analysis reveals wild and early released varieties have higher basal root angles, percentage of nodules, and number of lateral roots. Comparatively, root systems of more recently released lines are larger, with higher measurements for exploration parameters such as total root length, root convex area, and solidity. Correlations between root traits and leaf nutrition display that basal root angle, lateral root number, and nodule percentage have significant slightly positive correlations with leaf sulfur and potassium, while total root length, convex area, and solidity are significantly negatively correlated. The long-term impact of this research will inform breeders on how the selection process for aboveground plant architecture can affect root architecture and plant nutrition.
B-36
Selection of Superior Faba Bean Cultivars for the Mid-Atlantic Region of the United States
Frank Reith, Research Associate, University of Delaware
francisr@udel.edu
Co-authors: Emmalea Ernest, Rahul Raman, Maria Balota
Faba bean (Vicia faba L.) is a cool-season annual legume with worldwide cultivation. It may be grown as a fodder, grain, or vegetable crop. The species offers high protein content, 25-33%, substantial nitrogen fixation, 70-130 lb/ac, and can be grown for multiple markets. Several nations, such as Australia and Canada, have recognized the crop’s utility and have established breeding and research programs to support their growing faba industries. In the United States, the Mid-Atlantic and Southeastern states have insignificant cultivation of this crop. The University of Delaware, Virginia Tech, NC State, and the University of Maryland are working to research the viability of faba in our region. In 2023, a diversity panel (n = 515) representing available cultivars, experimental material, landraces, market collections, and all three subspecies from across the globe was compiled, primarily from the USDA germplasm collection. This panel was grown in Delaware and Virginia and sown in both the fall and spring of 2023-2024 and 2024- 2025. Pest and fertility management practices followed standard production recommendations, including irrigation at the Delaware site. Dry seed yield, seed weight and color, over-winter survival, stand and pod counts were phenotyped in all replications. The panel contains considerable variation for all phenotypes. Yield ranged from 0 – 1.2 tons per acre under winter sowing, to 0 – 1.6 tons per acre under spring sowing. While fall-sown plants needed to survive the winter to produce seed the following spring, cold tolerance had low correlation with yield or seed weight. Few vegetable types were winter hardy, and spring sowings had good succulent yields, at 14,500 lbs/ac. Many lines were poorly adapted to regional biotic and abiotic stressors, yet others show viability. Vegetable lines Windsor, Aquaculce, Vroma, PI 469172, PI 469170, and PI 510594 had agronomic and culinary value. Grain lines W6 17371, PI 557493, and PI 557493 demonstrate high yield and good agronomic qualities. These results highlight both the challenges and possibilities for faba bean production in the Mid-Atlantic U.S. Ongoing multi- year evaluations will allow identification of elite material for cultivar release and regional production recommendations, with the goal of supporting future faba adoption and industry development.
B-37
Genome-Wide Association Studies of Leaf and Pod Color Using the Snap Bean Association Panel (SnAP)
Burcu Celebioglu, Postdoc, University of California, Davis
bcelebioglu@ucdavis.edu
Co-authors: John P Hart, Timothy Porch, Phillip Griffiths, Travis A. Parker, and James R Myers
Color can be an indicator of plant health, quality, and productivity, and is useful to researchers to understand plant nutritional content in their studies. While the leaf color of green beans (Phaseolus vulgaris L.) supports the biomass accumulation and chlorophyll content of the plant, the color of its edible organ, the pod, affects the phytonutrient content. Although chlorophyll provides the base color, other compounds such as carotenoids and flavonoids may affect leaf and pod color. This research characterized leaf and pod color variation and identified significant loci in the 378-member SnAP. Phenotype traits were measured with a colorimeter using the L*a*b* scale while a GWAS was conducted to identify significant associated loci. Leaves were evaluated at three positions (lower, middle, and upper) in the canopy and both pod exterior and interior colors were obtained. The leaves at the upper level in the canopy were lighter than lower and middle-level leaves. Purple pods were darker (lowest L*) and yellow pods were lighter (highest L*). Green pods were generally two times higher for L* and lower in chroma (C*) compared with leaves. Broad sense heritability was low in pods, while narrow sense heritability was low in leaves. Moderate correlations between leaf L* and the interior and exterior pod L* imply that it would be possible to select for pod color based on leaf color, with verification using standard cultivars. More SNPs were associated with pod traits than with leaf traits. Fourteen SNPs with ≥10 PVE%, large SNP effect, and largest p-value for L* and H° pod exterior was identified on Pv01, Pv02, Pv03, and Pv08. The pod interior did not exhibit colors produced by anthocyanins or flavonols which allowed the differentiation of potential candidate genes associated with chloroplast and photosynthetic activity compared to the pod exterior where candidate genes related to both flavonoids and photosynthesis affected color. Several SNPs were associated with known qualitative genes including the wax pod locus (y), persistent color (pc), purple pods (V), and two genes expressed in seeds but not previously reported to affect other plant tissues (B and J).
B-38
Exploring Variation of Biological Nitrogen Fixation in the Common Bean Diversity Panel with Rhizobium Strains and Across Nitrogen Levels
Sachinthya Attanayake, Graduate Student, University of Saskatchewan
sachinthya.attanayake@usask.ca
Co-authors: K. E. Bett, I. J. Oreznik, I, G. C. diCenzo, and A. Vargas
Common bean (Phaseolus vulgaris), a vital global food and protein source and in general less efficient in biological nitrogen fixation (BNF) compared to other legumes. In Canada, the inconsistent performance of Rhizobium inoculants has limited adoption, hindering efforts to reduce reliance on synthetic nitrogen fertilizers. Enhancing BNF in beans is therefore critical for improving nitrogen use efficiency and supporting Canada’s sustainability goals. This study assessed the BNF potential of a 177 genotype Northern bean diversity panel (NBDP) inoculated with two Rhizobium strains collected in Canada (258 and 71) under contrasting nitrogen conditions. Beans were grown in sterilized vermiculite–sand medium under controlled conditions, and phenotyped for chlorophyll content (via normalized difference vegetation index, NDVI), nodulation, and biomass traits. A linear mixed model, with genotype as a random effect, showed that treatments had a strong effect on NDVI (F = 216.7, p < 2.2e-16). Fertilized controls (N+) exhibited the highest NDVI (mean 0.759), reflecting sufficient chlorophyll and nitrogen availability. Nitrogen-deficient plants (N–) displayed a pronounced decline in NDVI (–0.310, p < 2e-16), confirming reduced chlorophyll under stress. By contrast, plants inoculated with Rhizobium strains 258 and 71 maintained NDVI values statistically similar to N+ controls, indicating that inoculation preserved chlorophyll content at levels comparable to nitrogen-fertilized plants. Although genotypic variance in NDVI response was modest (σ² = 0.00018), it highlights genetic potential for improving chlorophyll maintenance under low-nitrogen conditions. Further analyses will integrate nodulation, biomass, and nitrogen content, with BNF quantified using the N-balance method. Genome-wide association studies (GWAS) will identify genetic loci and candidate genes linked to BNF-related traits. Overall, these findings suggest that Rhizobium inoculation can sustain chlorophyll content under nitrogen-limiting conditions, supporting genotype–strain combinations that improve nitrogen use efficiency and reduce fertilizer dependence in bean production.
B-39
Evaluation of Bean Fe Biofortification and Fe Bioavailability in East African Common Bean Germplasm
Raymond Glahn, Research Scientist, USDA
raymond.glahn@usda.gov
Co-authors: Winnyfred Amongi, Jason Wiesinger, Stanley Tamusange Nkalubo, Mildread Ochwo-Ssemakula, Badji Arfang, Thomas Lapaka Odongo, Ephraim Nuwamanya, Enoch Wembabazi, Phineas Tukamuhabwe, Karen Cichy, Clare Mukankusi
Since 2003, bean Fe biofortification (ie. high Fe content) has been one strategy to enhance the delivery of Fe in food systems and thus alleviate anemia. Biofortification is an approach that assumes a global average Fe concentration of 50 µg/g (dry weight) in non-biofortified beans and a minimally enriched threshold for Fe biofortification at 72 µg/g, with a highly biofortified bean at 90-94 µg/g, and assuming adequate Fe bioavailability. A second major strategy to enhance Fe delivery from beans is by enhanced Fe bioavailability; thus, identifying or selecting for factors and traits that promote Fe absorption while minimizing inhibitory components. Research over the past 5-10 years indicates that the assumptions of bean Fe biofortification are often not met, thus negating the benefit of the approach. At the same time, research has shown enhanced Fe bioavailability to be a more sustainable approach as it is based in seed coat genetics that control flavonoid production. The objective of this study was to evaluate Fe content and Fe bioavailability of 140 bean varieties, 39 of which were considered “biofortified,” meaning they should have Fe levels equal or greater than 72 µg/g. Iron content measurements showed that only 11 out of the 39 biofortified bean lines met the minimal levels of biofortification; whereas, 28 of the 101 non-biofortified bean varieties met or exceeded minimal biofortification. For iron bioavailability, varieties were compared to a white kidney (Snowdon) reference control known to have moderately high bioavailability. Eighteen varieties demonstrated equal or greater Fe bioavailability relative to the reference control, all of which were white beans. Of the remaining samples, approximately 4 yellows, 2 cranberry, 2 red-mottled and 1 purple variety exhibited moderate Fe bioavailability. Flavonoid analysis showed that the presence of procyanidins strongly inhibited Fe bioavailability. Most biofortifed lines (72%) did not contain enhanced levels of Fe. Higher Fe concentration does not consistently result in enhanced absorbable Fe. Iron bioavailability is defined by the genetics of seed coat color (ie. seed coat flavonoids). In conclusion, consistent with previous studies, Fe bioavailability should be the trait guiding breeding for enhanced Fe nutrition in beans.
B-40
Lima Seed Size and Color Classes Vary in Cooking Time, Iron Bioavailability, and Flavor/Taste Attributes
Jaclyn Adaskaveg, Postdoc, University of California, Davis
jaadaskaveg@ucdavis.edu
Co-authors: Jason Wiesinger, Yukina Murata, Antonia Palkovic, Raymond Glahn, Paul Gepts, Karen Cichy, Jean-Xavier Guinard, Christine Diepenbrock
Lima beans (Phaseolus lunatus L.) are nutrient-dense foods, rich in protein, fiber, vitamins, and minerals, but are underutilized in the U.S. White, large-seeded cultivars, such as UC92, or white, small-seeded cultivars, such as UC Haskell, are considered industry standards for dry limas sold in the U.S. However, germplasm collections show variation in many seed traits such as coloration/patterning, size, and shape—even among elite cultivars. Here, we characterized 36 lima entries including commercial cultivars and breeding lines for cooking time, macro- and micro-nutrients, and sensory attributes. Limas were categorized into baby- or large-seeded classes, which typically correspond to Mesoamerican and Andean domestications. Limas were also categorized by their white/green appearance or their various mottled seed coat colors, including orange, red and brown. Lima bean seeds were harvested from an on-station trial at UC Davis and sourced from commercial growers. Cooking time across lima entries was evaluated with a Mattson cooker. We found that large-seeded lima entries on average cooked faster than baby-seeded entries. We utilized near-infrared spectroscopy to predict the percentages of protein and fat content in seeds and found large-seeded types differed in grain macronutrient content compared to baby types. Seed concentrations of minerals including iron were evaluated with inductively coupled plasma-atomic emission spectroscopy. Flavonoids were quantified via LC-MS analysis. Bioavailable iron (iron available for absorption) was assessed via the Caco-2 Cell Bioassay. Limas with white/green seed coats had significantly higher iron bioavailability compared to lima beans with colored seed coats and a white navy common bean control, despite similar seed concentrations of iron. Further, among white/green seed types, large-seeded limas had significantly higher iron bioavailability than baby-seeded. Entries were also evaluated for sensory attributes (flavor, texture, appearance) through a descriptive analysis. We observed significant differences in flavor/taste attributes between white/green and colored lima beans, with entries having colorful seed coats being characterized as having bitter attributes while entries with white/green seed coats had sweet attributes. These findings highlight the potential to leverage variation in cooking time, nutritional and sensory properties of lima beans, potentially making them more appealing to consumers and stakeholders.
B-41
Diversity Assessment of Lima Bean Germplasm in the Clemson Breeding Program
Sangita Subedi, Graduate Student, Department of Plant and Environmental Sciences, Clemson University
sangits@g.clemson.edu
Co-author: Jenna Hershberger
Lima beans (Phaseolus lunatus) are a culturally significant vegetable in the southeastern United States. Moreover, lima beans are a nutritionally dense legume with nitrogen fixation ability and are adapted to diverse environments, which makes them a strong candidate for adaptation to rapid climate change. However, despite their agronomic and nutritional potential, the phenotypic and genotypic diversity in lima bean germplasm is still not well understood. Evaluation of this diversity will provide valuable insights for breeding programs, enabling the identification of potential parents to cross and develop improved lima bean varieties. Our project aims to evaluate a diverse panel of lima bean germplasm for morphological and agronomic traits across three years of field trials (2023-2025) at the Clemson PDREC in Florence, SC, with an augmented incomplete block design. Morphological traits, including leaf shape, seed, and pod characteristics, were measured through image analysis, and the nutrient content of succulent beans was quantified through laboratory analysis. Growing degree days (GDD) for each accession at 50% flowering and pod maturity were calculated using daily minimum and maximum temperatures. Phenotypic data were analyzed using principal component analysis and correlations in R. Results show wide variation among accessions for all traits, including GDD to 50% flowering (379 - 1006), GDD to pod maturity (372 - 698), and total protein (21 % - 32 %). These phenotypic traits will be linked to genetic data to identify their genetic basis through genome-wide association studies. Phenotypic data will be uploaded to GRIN, making it public.
B-42
High-Throughput Phenotyping of Canning Quality Traits in Dry Beans Using Computer Vision and Deep Learning
Lovepreet Singh, Postdoc, Michigan State University
singhlo2@msu.edu
Co-authors: Om Sai Madhav Lella, Evan Wright, Valerio Hoyos-Villegas
Canning quality assessment in dry bean breeding programs remains a major challenge, with color retention being the key trait influencing consumer acceptance and commercial value. Traditional approaches of color evaluation rely on either subjective scoring or expensive colorimetry instruments. These methods have significant drawbacks, including perception bias and subjectivity in panelist ratings, as well as the high cost of instruments and their inability to capture spatial information. This study aimed to develop a high-throughput, automated, and objective phenotyping pipeline that integrates computer vision and deep learning technologies to improve canning quality assessment across diverse market classes in dry beans. This pipeline combines the YOLOv8n object detection model with the segment anything model (SAM) for precise bean segmentation. The pipeline then extracts pixel-level CIE L*a*b color values from segmented images, applies calibration using partial least squares (PLS) regression, and computes an objective distance-based D-Score, which quantifies the Euclidean distance from an ideal reference color. We evaluated 1,300 dry bean entries from Michigan State University and USDA-ARS across four years using this pipeline. The YOLOv8n model achieved near-perfect detection performance across 13 diverse market classes, with precision and recall values approaching 1.0. Similarly, the segment anything model achieved almost 100% segmentation accuracy. Comparative analysis showed that the image-derived D-Score outperformed panelist ratings and colorimetry methods, with significantly lower prediction error in regression models compared to traditional models. The D-Score metric offered high resolution in color assessment, enabling the distinction of subtle, genotype-level differences while traditional methods were unable. Principal component analysis revealed that lightness drives 96.8% of color quality variation, with chromaticity components providing refinement to the D-Score. Additionally, the D-Score provides a ranking criterion that enables breeders to make informed decisions and select superior genotypes. This pipeline was deployed to MSU’s High Performance Computing Center (HPCC) and processes 200 high-resolution images in under five minutes, making it practical for large-scale breeding applications while eliminating subjective bias and reducing costs. This open-source phenotyping pipeline provides breeding programs with an accessible, objective, and scalable tool for accelerating genetic gain in canning quality traits across diverse bean market classes.
B-43
Comparison of Selection Strategies to Increase Yield Gain in Dry Bean
Maryam Vazin, University of Guelph
Co-author: K. Peter Pauls
An understanding of the genetic bases of complex traits, such as yield in crops like common bean (Phaseolus vulgaris L.) is important for designing sustainable strategies for addressing growing global food demands, in the near future. However, yield and its related traits are controlled by multiple with major and minor alle effects. To determine what selection strategy should be implemented to optimize genetic gain for yield in common bean population a Nested Association Mapping (NAM) population of F4:5 recombinant inbred lines (RILs) was created with the cultivar Ex Rico 23, and 10 founder lines that span the genetic diversity of Ontario Mesoamerican germplasm. The NAM population was evaluated for different agronomic traits including yield, days to 50% flowering, and days to maturity in the field in four environments. The distributions for all the traits (days to 50% flowering, days to maturity, yield, and yield/day) were continuous and included transgressive segregations for each environment. In addition, the susceptibility and resistance to Anthracnose, an important biotic stress in bean, was evaluated in the greenhouse for one subpopulation. A comparison of the relative effectiveness of phenotypic selection, genomic selection and QTL index selection will be presented.