Organ-Differentiated Metabolomic Profiling Reveals Putative Associations Between Predicted Bioactivity and Database-Annotated Flavor in Medicinal–Edible Talinum paniculatum (Jacq.) Gaertn

Abstract

Background: Medicinal–edible plants are abundant sources of natural bioactives and functional food precursors. Talinum paniculatum (Jacq.) Gaertn is a commonly utilized folk herb, but there is still a lack of systematic metabolomic research linking its organ-specific phytochemical variations to its medicinal and sensory attributes. Methodology: In this work, untargeted UPLC-MS/MS metabolomics was applied to methanolic extracts of the roots, stems, leaves and fruits of T. paniculatum. Multivariate statistics and KEGG enrichment screened differentially accumulated metabolites (DAMs), while multi-database matching against TCMSP, FlavorDB, and MFood-Flavor Innovation was performed to prioritize candidates with database-assigned pharmacological or flavor-related annotations. For candidates carrying flavor descriptors, relative abundance values were compared across organs to characterize organ-dependent distribution patterns. Results: After quality control filtering and annotation-based deduplication, 4521 putatively annotated metabolite entries spanning 21 chemical classes were retained across the four organs. Amino acid derivatives constituted the largest class, accounting for 31.72% of all annotated entries. Among the presumptively annotated DAMs, five were further proposed as candidate organ-discriminatory metabolites. TCMSP screening matched 348 metabolites with predicted pharmacological relevance; 55 met the OB and DL screening criteria, and of these, 22 showed preferential accumulation in fruits. Subsequent database annotation highlighted candidates carrying both predicted pharmacological relevance and database-assigned flavor attributes, including umami-associated pyroglutamic acid, sweet neoastilbin, and bitter-tasting melilotoside, helenalin, quercetin, cianidanol and glycitein. These dual-functional metabolites exhibited distinct organ-specific distribution patterns. Conclusions: The organ partitioned distribution of metabolites with database-assigned flavor attributes and predicted bioactivity provides testable hypotheses for selective organ utilization, pending targeted quantification, sensory evaluation, and bioactivity validation.