A high-throughput strategy for screening umami peptides from Chouguiyu based on peptidomics and machine learning-assisted molecular simulation.

To develop a high-throughput strategy for screening umami peptides, peptidomics was integrated with machine learning-assisted virtual screening, molecular docking, molecular dynamics simulations, and electronic tongue evaluation. Peptidomics analysis of Chouguiyu revealed peptide profiles, from which eight candidate umami peptides (SGLVKAGFAGDDAPR, EAVGDWRKNIED, NGSGLVKAGFAGDD, SGGTTMYPGIADR, DNGSGLVKAGFAGD, AGIHETAYNSI, AGIHETAYN, and TGDGVNDAPALK) were identified and chemically synthesized. Molecular simulation results demonstrated that all eight peptides showed stable interactions with the umami receptor T1R1/T1R3, with key binding residues including Asp219, Asp218, and Glu301, primarily mediated through hydrogen bonding interactions. Electronic tongue analysis confirmed that all peptides exhibited umami characteristics, with intensity values more than twice that of a 0.1% MSG reference solution. These findings provide new insights into the molecular basis of umami perception and offer an efficient strategy for the discovery of taste-active peptides in fermented foods.