Taste-active peptides, end-to-end: from generation and sensory validation to AI-driven discovery

Taste-active peptides (TAPs) are food-protein-derived peptides that elicit or modulate gustatory sensations; presently gaining interest as natural umami-enhancing, bitterness-modifying, kokumi, and saltiness-aiding molecules. This review provides a synthesis of progress across discovery pipeline steps, from hydrolysis-based generation and process-intensified extraction to purification/identification workflow and sensory interpretation. Receptor-level mechanisms involved in peptide taste perception are summarized, and structure-activity patterns associated with important taste modalities are collated. On the data side, computational models are profiled for prioritization of candidates from proteome-derived libraries, with emphasis on curated peptide-receptor labels, transparent feature construction (e.g., sequence/physicochemical descriptors with docking-derived interaction summaries), and reproducible benchmarking. Practical bottlenecks (dataset curation, assay comparability in sensory validation, and scalable computation) are mapped together with emerging resources that can standardize tasks and reporting. By combining approaches and demonstrations from experimental and computational disciplines, the review offers an operations-focused reference for the design and validation of TAPs in clean-label product development.