A multi‐environment GWAS trial reveals novel loci for yield traits in a MAGIC population of Upland cotton ( Gossypium hirsutum L.)

Abstract The genetic architecture underlying yield traits is not fully understood. To dissect the genetic basis of lint yield (LY) and its components, we conducted a genome‐wide association study (GWAS) using a multi‐parent advanced generation intercross (MAGIC) population consisting of 372 recombinant inbred lines (RILs). This population was developed from 16 parents, including 12 exotic germplasm lines derived from five tetraploid Gossypium species and four elite Upland cotton cultivars. The population underwent five generations of random mating and five generations of self‐pollination. The RILs were evaluated at three locations in the US Cotton Belt over 8 years. The MAGIC population displayed greater phenotypic variability than the parental lines, with all traits showing transgressive segregation. When GWAS was performed across environments, a total of 244 significant single nucleotide polymorphisms (SNPs) were identified for LY and related traits. Among them, 75 SNPs were novel. Twelve stable SNPs were identified across different environments and genetic backgrounds, including four for lint index (LI), three for seeds per boll, and two each for LY, lint percent (LP), boll weight, and seed index. In addition, four SNPs exhibited pleiotropic haplotype effects on LY, LP, and LI, consistent with the observed phenotypic correlations among these traits. For yield, the candidate gene ( Gh_A08G0526 ) known for its role in fiber initiation, elongation, and overall yield was located within the previously identified quantitative trait locus region on A08 at 107 Mb. The discovery of novel and stable pleiotropic loci advances our understanding of the genetic architecture of yield traits.