Integrated multi-omics analysis of flavonoid metabolism in soybean seeds of different colours reveals candidate regulatory networks associated with anthocyanin and isoflavone accumulation

Soybean seed colour reflects divergence in flavonoid metabolism and is an important quality attribute for edamame and functional food ingredients. The transcript-metabolite patterns associated with anthocyanin and isoflavone accumulation across colour phenotypes remain incompletely characterised. We compared one black-seeded (B), one green-seeded (G), and one yellow-seeded (Y) accession using commercial colorimetric assays, widely targeted metabolomic profiling, and RNA sequencing. Whole-seed assays showed higher total anthocyanin contents in B and G than in Y and a higher total flavonoid content in B than in G and Y. B also showed higher relative abundances of several delphinidin derivatives, proanthocyanidins, and catechins. Selected cyanidin and pelargonidin derivatives were more abundant in Y, whereas daidzein and several triterpenoid saponins were more abundant in G. Transcriptome comparisons identified accession-associated differences in flavonoid-pathway genes, including members of the CHS, CHI, and DFR families. Several flavonoid-related transcripts and metabolites showed parallel patterns across the three accessions. Exploratory DIABLO analysis revealed coordinated variation among the phenotype, metabolite, pathway-gene, and transcription-factor datasets and retained 20 transcription factors across two model components. RT-qPCR mean patterns were broadly similar to the RNA-seq results for some selected genes. Together, these findings reveal distinct flavonoid profiles and associated transcriptional patterns among the three soybean accessions, providing candidate metabolites, genes, and transcription factors for further functional and biochemical investigation.