High-Density Genome-Wide Association Mapping Identifies Candidate Loci Associated with Maize Stalk Cell Wall...

Maize (Zea mays L.) stalk cell wall composition is a key determinant of forage digestibility, lodging resistance, and biomass utilization efficiency. Although previous genome-wide association studies (GWAS) have identified loci associated with lignin (LIG), cellulose (CEL), and hemicellulose (HC), advances in genomic resources provide an opportunity to revisit existing phenotypic datasets at substantially higher resolution. Here, we re-analyzed a maize association panel consisting of 341 diverse inbred lines using an expanded genotype dataset containing 10.77 million SNPs, two derived compositional indices (CEL/HC and [LIG/(CEL + HC)], and six complementary GWAS models. Across all traits and models, we identified 855 unique significant SNPs associated with 579 candidate genes. Among the traits examined, LIG/(CEL + HC) yielded the greatest number of associations, suggesting that indices representing the relative balance among cell wall components may better capture the genetic architecture of cell wall composition than individual component measurements alone. Integration of multiple GWAS models with functional enrichment, haplotype, and selective sweep analyses prioritized three biologically relevant candidate genes encoding a MYB58 transcription factor, the glycosyltransferase Xt9, and a putative xyloglucan 6-xylosyltransferase. Haplotype analysis revealed significant effects of Xt9 and the xyloglucan 6-xylosyltransferase on cell wall composition, while selective sweep analysis identified Xt9 as a target of repeated selection during maize domestication, ecological adaptation, and modern breeding. Although these candidate genes provide promising targets for future investigation, the associations identified here are based on a single association panel and require functional and independent population validation. Collectively, our results demonstrate how high-density genotyping combined with complementary GWAS models can refine candidate associations and generate testable hypotheses from existing phenotypic datasets.