Maize association mapping panel

The comprehensive dataset of metabolites offers an opportunity to gain insights into the metabolic divergence during the domestication and improvement of maize. Partial least squares discriminant analysis (PLS-DA) was employed to identify the metabolites that significantly contribute to subgroup divergence. Our results indicated that 43.5% and 47.2% of the metabolite features are associated with the divergence in Field-Waxy and Field-Sweet, respectively, as evidenced by variable importance in projection (VIP) scores exceeding 0.8. Even with stringent criteria, with VIP scores exceeding 1, 28.9% and 34.0% of the metabolites were found to be significant in Field-Waxy and Field-Sweet, respectively. Furthermore, the paired t-test (adjusted for false discovery rate (FDR) < 0.05) or fold change (|Fold Change| ≥ 2 or 1.2) analysis of metabolite levels revealed that a majority of these metabolite features exhibit noteworthy divergence or significant changes.

We conducted a comprehensive genome-wide association study (GWAS) using a linear mixed model (LMM) on a set of 20.3Mb credible single nucleotide polymorphisms (SNPs). To determine significant associations between metabolic traits and genetic variations, we applied a Bonferroni correction with a threshold of P = 4.63 × 10-8. However, it is important to note that the presence of significant genetic differences between sweet corns and other corn varieties (such as waxy and field corns) might obscure the effects of loci with underlying contributions. To overcome this limitation, we identified significant signals in sweet corns, other corn varieties, and all accessions to obtain a more comprehensive understanding of the associations. Remarkably, nearly all of the detected metabolites (755 out of 802) exhibited at least one significant association with genetic variations, with an average of 8.36 associations per metabolite. Furthermore, more than 42% (337 out of 755) of the metabolites displayed no more than 5 associations, indicating that the majority of metabolites are regulated by a small number of genes. To further investigate these candidate genes, we performed haplotype analysis for each gene.