Recent advances and applications of deep learning, electroencephalography, and modern analysis techniques in screening, evaluation, and mechanistic analysis of taste peptides

Abstract

Background

Taste peptides are oligopeptides that improve the flavor and palatability of food. Due to their unique taste characteristics and nutritional values, the development of taste peptides has become a hot spot for food flavoring research and commercial applications. Screening and evaluating of taste peptides based on traditional experimental methods is inefficient and labor-intensive. Deep learning and electroencephalogram can enable high-throughput screening of taste peptides and analysis of taste mechanism, which has attracted extensive attention.

Scope and approach

This review summarizes the structural characteristics, taste cells and their subtypes, as well as cellular mechanism of taste transduction of the taste peptides. Significant attention has been focused on the high-throughput screening of taste peptides using deep learning model. Furthermore, the methods for evaluating taste intensity and clarifying the taste mechanism of taste peptides have also been reviewed.

Key findings and conclusions

The application of deep learning in the high-throughput screening of taste peptides maintained a strong prediction performance. Notably, the combination of multiple deep learning algorithms could enhance the accuracy of predicting taste peptides compared to a single algorithm. In addition, the application of electroencephalogram, bioelectronic tongue, and taste organoids-on-a-chip is an effective way to evaluate taste intensity, while fluorescence spectrum, surface plasmon resonance, and molecular docking are suitable for elucidating the taste-presenting mechanism of taste peptides. The review can provide a valuable reference for high-throughput screening of taste peptides and increases its application in the condiment industry.