Artificial intelligence-assisted screening has identified six fungal-derived peptides that can produce a salty taste or enhance saltiness, providing molecular candidates for developing sodium-reduced foods. The study combines in silico protein digestion, embedding-based machine learning, electronic-tongue measurements, cellular assays, molecular docking and molecular-dynamics simulations. It was published as an accepted article in npj Science of Food and presents the peptides as promising ingredients rather than established commercial solutions or proven nutritional interventions.
The researchers first built a dataset of taste-active peptides and compared three molecular-embedding approaches: Uni-Mol, ESM-3 and MolCLR. All three were evaluated with the same FT-Transformer classification framework. ESM-3 gave the most favorable classification performance and was selected to screen peptides generated through in silico hydrolysis of proteins from 14 fungal proteomes. Six candidates were then synthesized for experimental assessment. Three were classified as salty peptides—AGHQR, AHDQY and SKPGNR—and three as saltiness-enhancing peptides: IPVVVDL, DNNEEW and ANSDL.
Electronic-tongue analysis distinguished the most effective candidates under the laboratory conditions tested. AHDQY was the strongest salty peptide in deionized water, while ANSDL was the most effective saltiness enhancer in a low-salt sodium chloride system. These measurements show that the screened compounds could modify saltiness in the specified test solutions. The reported results do not establish the amount of sodium that could be replaced in finished foods, whether the ingredients could perform similarly across different formulations, or whether their use would produce health benefits in people.
In cellular experiments, the peptides exhibited good cytocompatibility and, to varying degrees, protected human umbilical vein endothelial cells from high-salt-induced injury. That protection was accompanied by restored nitric oxide production and reduced accumulation of reactive oxygen species. These findings concern responses in an experimental cell model and are not evidence of cardiovascular benefits in humans. Molecular docking and molecular-dynamics simulations also supported stable interactions between the selected peptides and transmembrane channel-like protein 4, together with peptide-associated changes in local receptor flexibility and secondary structure. The combined screening, sensory, cellular and computational evidence identifies AHDQY and ANSDL as leading candidates for further investigation of natural saltiness-modulating ingredients for sodium-reduced foods.
Subject of Research: AI-guided screening of fungal peptides for saltiness modulation in sodium-reduced foods
Article Title: AI-guided discovery of fungal-derived salty and saltiness-enhancing peptides for sodium-reduced foods
Article References: Zhang, Y., Shen, X., Liu, Y., Li, W., & Han, W. (2026). AI-guided discovery of fungal-derived salty and saltiness-enhancing peptides for sodium-reduced foods. npj Science of Food. https://doi.org/10.1038/s41538-026-01209-4
Image Credits: AI Generated
DOI: 10.1038/s41538-026-01209-4
Keywords: artificial intelligence, fungal peptides, saltiness enhancement, sodium reduction, food science, machine learning, electronic tongue, cytocompatibility, molecular docking, protein embeddings, AI-guided, discovery
News Source: Blake Davidson. (October 20, 2026). AI-guided discovery of fungal-derived salty and saltiness-enhancing peptides for sodium-reduced foods. Scienmag.
