Genome-wide association study reveals candidate genes controlling root system architecture in pepper (Capsicum annuum L.)

• Root system architecture (RSA) of 137 pepper ( Capsicum annuum L.) accessions was quantified from images of 4-week-old seedlings grown in a paper-based medium. • A genome-wide association study (GWAS) identified four candidate genes associated with three significant SNPs on chromosomes 4, 6, and 9. • Correlation analysis between root and aboveground traits revealed positive associations between RSA traits and first internode length in 2-month-old plants. • Genomic prediction (GP) using GWAS-derived SNPs achieved a prediction accuracy of 0.63 for total root length (TRL) in pepper. Although plant growth habits play a pivotal role in horticultural productivity, research has predominantly concentrated on aboveground organs, with root system architecture (RSA) frequently overlooked due to the inherent difficulty of accessing subterranean structures in situ . In this study, we quantified RSA in 137 pepper ( Capsicum annuum L.) accessions by analyzing images of four-week-old seedlings grown in a paper-based medium, measuring primary root length (PRL), sum of lateral roots (SLR), number of lateral roots (NLR), and total root length (TRL) in 137 accessions. A genome-wide association study (GWAS) identified three significant SNPs for RSA on chromosomes 4, 6, and 9. Linkage disequilibrium and epistasis analyses identified candidate genes associated with RSA, including Root System-Inducible 1 ( RSI-1 ), Small-Intrinsic Protein 1-2 ( SIP1-2 ), and G-type lectin S-receptor-like serine/threonine protein kinase ( SRK ). We further evaluated associations between RSA and aboveground traits by analyzing growth-related traits at 2 and 6 months, as well as RGB image–based drought tolerance indices. Correlation analyses revealed positive associations between RSA and the 2-month traits, first internode length, whereas associations with 6-month traits and drought indices were weak or non-significant. To utilize GWAS data, we performed genomic prediction (GP) using GWAS-derived SNPs and achieved a prediction accuracy of 0.63, indicating moderate predictive power for TRL. Taken together, these findings provide genetic insights into root development in pepper and establish a genetic framework for future efforts to incorporate RSA traits into breeding programs.