From seed shape to genetics: analyzing medicago spp. seed morphology using gwas and machine learning

Authors: BOTKIN, JACOB - University Of Minnesota; MEDINA, CESAR - University Of Minnesota; Park, Sunchung - Sun; POUDEL, KABITA - University Of Minnesota; CHA, MINHYEOK - Korea University; LEE, YOONJUNG - University Of Minnesota; Prom, Louis; Curtin, Shaun; Xu, Zhanyou; Ahn, Ezekiel

Submitted to: Scientific Reports
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/14/2024
Publication Date: 7/30/2024
Citation: Botkin, J.R., Medina, C.A., Park, S., Poudel, K., Cha, M., Lee, Y., Prom, L.K., Curtin, S.J., Xu, Z., Ahn, E.J. 2024. From seed shape to genetics: analyzing medicago spp. seed morphology using gwas and machine learning. Scientific Reports. 14:17588. https://doi.org/10.1038/s41598-024-67790-4.
DOI: https://doi.org/10.1038/s41598-024-67790-4

Interpretive Summary: Alfalfa (Medicago sativa L.) is a perennial forage cash crop with extensive economic and agronomic importance due to its high nutritional value. An essential part of alfalfa research includes the production of high-quality alfalfa seeds, but the fundamental genetic architecture of seed morphology in alfalfa is relatively unexplored. This study investigated the diverse morphological variations in alfalfa seeds, focusing on size, shape, and color. With a combination of robust statistical analysis and machine learning models, candidate genes potentially confer significant morphological differences found in this study, providing valuable insights into the underlying genetic diversity, which can be applied in breeding strategies to develop improved alfalfa cultivars with desirable agronomic traits.

Technical Abstract: Plant seed morphology is shaped by multiple components including genetic, physiological, and ecological components that highly affect yield, quality, and market price. The importance of Alfalfa (Medicago sativa L.) is often emphasized by calling it the "Queen of the forages." Given this great importance, understanding the seed morphology of alfalfa was imperative. To unravel the morphological characteristics and elucidate the genetic bases of seed morphology in alfalfa, we screened 318 lines, including 244 lines of alfalfa and 28 other lines of alfalfa relatives, for 11 seed morphologies associated with size, shape, and color. Statistical analysis was applied to identify potential relationships among the 11 morphological traits. Based on the phenotypic data of M. sativa subsp. sativa, a Genome-wide association study (GWAS) was conducted by using publicly available single nucleotide polymorphisms (SNPs). When mapped to the reference Medicago truncatula genome, multiple candidate genes potentially associated with alfalfa seed morphology were identified (threshold of 1E-04). Genes in proximity to top SNPs associated with seed morphology traits included CPR1, MON1, a PPR protein, and Wun1. Machine learning models were utilized to validate the GWAS results, as well as identify additional marker-trait associations across the whole genome and within coding sequences. Along with the strong correlations observed between the seed morphological traits, these candidate genes will play important roles in understanding the genetic basis of seed morphology in Medicago spp.