Food-Derived Umami Peptides: Linking Candidate Discovery, Experimental Validation, and Food Applications

Abstract

Alternative proteins and cultivated meat increase the need for flavour strategies that maintain savoury quality with less reliance on salt and conventional seasonings. Food-derived umami peptides are promising candidates, but their practical value depends on evidence beyond sequence prediction. With advances in virtual digestion, peptidomics, and AI-assisted approaches, discovery is expanding from sensory-guided preparation toward candidate-library construction and data-driven prioritization. This review compares food-protein resources and experimental and sequence-based discovery routes, and then examines chemical and sensory validation, instrumental support, and receptor-related evidence. Structure-based calculations are considered tools for mechanistic interpretation rather than substitutes for functional assays. The review organizes these findings within a prediction–validation–application framework and assesses their relevance to food formulation, processing stability, and production feasibility. Remaining priorities include transparent taste datasets, reproducible model evaluation, and replicated food trials that connect effective dose, off-flavours, recovery yield, and cost. A data–model–experiment–food framework is proposed to guide practical umami-peptide development.