Feasibility and pitfalls of combining small subpopulations for genome‐wide association analysis of maize (Zea mays L.) kernel traits

Abstract

Abstract Genome‐wide association studies (GWASs) typically require large populations to achieve sufficient statistical power; however, maintaining large populations is labor‐intensive and costly. In this study, we evaluated the feasibility of improving GWAS detection power by strategically integrating small maize (Zea mays L.) subpopulations for kernel‐related traits. Using high‐throughput phenotyping of eight kernel traits, GWAS was conducted across four different population combinations. Expanding population size generally increased single‐nucleotide polymorphism (SNP) association strength and enabled identification of additional loci, supporting the advantage of subpopulation integration. However, some loci identified in smaller combinations lost association strength after incorporating genetically divergent subpopulations, indicating that phenotypic distribution and genetic background strongly influence GWAS outcomes. Comparison with previously reported kernel quantitative trait loci revealed that many identified SNPs reside in multi‐trait genomic regions, while several loci exhibited trait‐specific effects. Candidate genes within linkage disequilibrium blocks included pentatricopeptide repeat proteins, an insulin‐degrading enzyme‐like protein, and a peptidyl‐prolyl cis–trans isomerase, all with plausible roles in kernel development. Collectively, our results suggest that integrating small populations is an effective and practical strategy to enhance GWAS feasibility and locus discovery, while careful consideration of population structure remains essential.