Multi-omics profiling reveals developmental regulation of lipids and bioactive metabolites in Akebia trifoliata seeds

Akebia trifoliata is an emerging oilseed crop with significant potential for functional foods due to its unique profile of lipids and bioactive compounds. A. trifoliata seeds also undergo a dramatic color transition during maturation, accompanied by extensive metabolic reprogramming. However, the coordinated regulatory landscape linking gene expression to this dynamic metabolome remains poorly understood. To elucidate the molecular mechanisms governing this process, we performed integrative transcriptomic and metabolomic analyses across four developmental stages. Metabolomic profiling identified 2134 distinct metabolites, revealing significant stage-specific alterations. The early SA stage was characterized by upregulation of lipids, flavonoids, and alkaloids, concomitant with high expression of genes such as CER1, ERD14, and CHS. In contrast, later stages (SC to SD) exhibited increased accumulation of lipids and flavonols. Transcriptome analysis identified 6779 differentially expressed genes (DEGs) between SA and SD stages, with significant enrichment in secondary metabolism, pigment biosynthesis, and plant hormone signaling pathways. At the terminal SD stage, protective genes such as AKT017106 (AMP1) and AKT016065 (HSP20) were highly expressed. Fuzzy C-Means (FCM) clustering of both datasets revealed coordinated temporal patterns. Integrated correlation analysis confirmed strong co-regulation, linking early-stage gene clusters (enriched in hormone signaling) to the accumulation of phenolic acids and flavonoids. Conversely, late-stage gene clusters involved in stress response correlated with lipid and amino acid accumulation. By combining FCM clustering with systematic cluster-cluster correlation analysis, an approach not previously applied to A. trifoliata seed development, this study organizes transcriptome-metabolome relationships into a coherent temporal framework. It nominates a defined set of candidate genes and metabolites for future functional validation and breeding efforts, providing a hypothesis-generating resource rather than a validated regulatory network.