The study shows that maize plants carrying autophagy-defective atg10 mutations exhibit delayed flowering and significant reductions in kernel size, weight, and number, culminating in lower grain yield. Reciprocal crossing experiments reveal that the maternal genotype, rather than the seed genotype, primarily drives the observed kernel defects, suggesting impaired nutrient remobilization from maternal tissues during seed development.

The study evaluated whether integrating genomic, transcriptomic, and drone-derived phenomic data improves prediction of 129 maize traits across nine environments, using both linear (rrBLUP) and nonlinear (SVR) models. Multi-omics models consistently outperformed single-omics models, with transcriptomic data especially enhancing cross‑environment predictions and capturing genotype‑by‑environment interactions. The results highlight the added value of combining transcriptomics and phenomics with genotypes for more accurate and generalizable trait prediction in maize.

The study generated a temporal physiological and metabolomic map of leaf senescence in diverse maize inbred lines differing in stay‑green phenotype, identifying 84 metabolites associated with senescence and distinct metabolic signatures between stay‑green and non‑stay‑green lines. Integration of metabolite data with genomic information uncovered 56 candidate genes, and reverse‑genetic validation in maize and Arabidopsis demonstrated conserved roles for phenylpropanoids such as naringenin chalcone and eriodictyol in regulating senescence.