Plant Breeding Seminar Series:
Wednesday, April 8^th, 2026 Agronomy 3140, 3:00-4:00pm

“Pixels to Predictions: Leveraging deep learning for high-throughput phenotyping to uncover the genetic architecture of mungbean pod and seed traits.”

Venkata Naresh Boddepalli

Ph.D. candidate in plant breeding, Dr. Arti Singh’s Lab, Department of Agronomy, ISU

Seminar Description: Improving seed yield and quality is a primary breeding goal for mungbean (Vigna radiata (L.) R. Wilczek), yet progress has been historically constrained by labor-intensive manual phenotyping and limited genomic resources. This study presents an integrated framework that combines deep learning-based phenomics with multi-model genome-wide association studies (GWAS) and genomic prediction (GP) to dissect the genetic architecture of pod and seed traits in the Iowa Mungbean Diversity (IMD) panel. We deployed an automated pipeline utilizing the Segment Anything Model (SAM) and DeepLabV3 to extract precise morphological data from over 3000 scanned images. The digital measurements demonstrated exceptional reliability, achieving high correlations with manual ground-truth data (r > 0.96 for pod dimensions and r > 0.91 for seed dimensions. Using these high-fidelity phenotypes, a multi-model GWAS identified 79 unique SNPs for seed traits and 65 unique SNPs for pod traits, revealing significant genomic hotspots on chromosomes 4, 6, and 8. Notably, we identified several high-confidence pleiotropic loci, including two major-effect SNPs on 4 (5_35265704) and chromosomes 6 (5_21173402), which explained over 20% of the phenotypic variance for pod and seed size traits. Comparative mapping across legumes and Arabidopsis identified conserved candidate genes involved in auxin signaling (GH3 family), transmembrane transport (ABC transporters), and potassium homeostasis (HAK5), suggesting an evolutionarily conserved network governing legume reproductive development. Furthermore, genomic prediction models yielded high accuracies for pod and size-related traits (r = 0.81–0.86) with up to 80% selection coincidence index, demonstrating the potential of genomic selection to accelerate genetic gains. Collectively, this research provides a scalable, AI-driven toolkit to modernize mungbean breeding, offering validated markers and predictive models to improve productivity and market quality in the United States and globally

Venkata Naresh Boddepalli is a fourth-year Ph.D. student in Plant Breeding at Iowa State University, specializing in genomics and AI-assisted phenomics under the supervision of Dr. Arti Singh. His research focuses on integrating advanced data analytics and high-throughput phenotyping to modernize mungbean crop improvement, including the development of AI-enabled imaging systems that have reduced manual scoring time by 60%. Prior to his doctoral studies, Naresh spent 8 years as a Scientific Officer at the World Vegetable Center (WorldVeg) in India, where he led multi-national legume breeding projects across South/Central Asia, Africa, and Australia and trained several researchers in mungbean breeding and digital database management. He holds an M.Sc. in horticulture with a focus on vegetable breeding and has contributed to over 15 peer-reviewed publications and conference papers. Currently serving as the President and Co-chair of the 13th Raymond F. Baker Plant Breeding Symposium, Naresh continues to bridge the gap between traditional breeding and cutting-edge digital transformation.

“Impact of spontaneous haploid genome doubling on haploid induction in maize haploid inducers”

 Vencke Grüning

 PhD candidate in Plant Breeding, Dr. Thomas Lübberstedt, Department of Agronomy, ISU

Seminar Description: Doubled haploid (DH) technology is a cornerstone of modern maize breeding, enabling rapid development of homozygous inbred lines. Haploid inducer pollen generates haploid seeds on donor plants, but haploids are also produced during self-pollination for inducer seed increase. As haploid induction rates (HIR) increase, the number of self-induced haploid plants which are typically smaller and male sterile, also increases. While haploid seedlings used for DH line development are treated with for example colchicine to induce genome doubling, inducer seeds for production fields are directly sown without chemical treatment. Incorporating spontaneous haploid genome doubling (SHGD) into inducer lines could restore fertility in self-induced haploids and reduce the cost and complexity of DH production by eliminating chemical treatments and transplanting.

To evaluate whether haploid male fertility (HMF), conferred by SHGD, and HIR are independent traits, a haploid population segregating for qshgd1 (SHGD) and mtl (HIR) was developed. Seedlings were divided into colchicine-treated and untreated groups and transplanted to the field. Approximately 1,000 haploid plants were genotyped using functional markers, and HIR was measured through testcrosses from the doubled haploid population. Preliminary results confirmed segregation distortion against mtl and higher pollination success in plants carrying qshgd1, with no significant association between SHGD and HIR.

 Vencke Grüning is a PhD candidate in Plant Breeding at Iowa State University under the supervision of Dr. Thomas Lübberstedt. She received her Bachelor’s and Master’s degrees in Crop Science from the University of Göttingen, Germany. Prior to her PhD, she worked in France at Limagrain on a breeding research project. Her current research focuses on improving the efficiency of the DH technology in maize.

“Vernalization Requirements and Freezing Tolerance in Poa secunda: Insights for Perennial Ground Cover Systems”

 Vipul Kumar

 PhD student, Dr. Shuizhang Fei, Department of Horticulture, ISU

Seminar Description: Perennial ground cover systems are being explored as a sustainable approach to improve soil health, reduce erosion, and enhance ecosystem services in row-crop agriculture. Poa secunda, a cool-season perennial grass native to North America, is a potential candidate for these systems due to its vegetative summer dormancy trait. This seminar will present research examining the vernalization requirements and freezing tolerance of three Poa secunda cultivars to better understand their adaptation to temperate environments. Understanding how these plants respond to cold exposure and flowering cues is important for evaluating their suitability as perennial ground covers in cropping systems. Additionally, preliminary insights into the reproductive biology of Poa secunda will be discussed using Flow Cytometry Seed Screen (FCSS) to evaluate the degree of apomixis. These findings will contribute to identifying germplasm with traits suitable for resilient perennial ground cover systems.

Vipul Kumar is a Ph.D. student in the Fei lab in the Department of Horticulture at Iowa State University. His research focuses on understanding adaptive traits in Poa secunda, including summer vegetative dormancy, vernalization requirements, freezing tolerance, and reproductive biology, to support the development of perennial ground cover systems. His work integrates physiological experiments with genomic approaches, including RNA sequencing, to identify genetic mechanisms underlying vegetative summer dormancy. Vipul received his bachelor’s degree in agriculture and his master’s degree in genetics and plant breeding from CCS Haryana Agricultural University, India.