INFORMATION TECHNOLOGY

Machine learning can effectively screen for “bitterness”


Bitterness is often not popular in the taste of food, and one of the important factors that cause this bitterness is a biomolecule – bitter peptide. Bitter peptides are small peptides that bind to bitter receptors on cell membranes to trigger bitter taste perception and are typically produced during food processing, storage, or digestion. Recently, the team of Jin Yan, a researcher at the Dalian Institute of Chemical Physics, Chinese Academy of Sciences, cooperated with Dalian Polytechnic University and Inner Mongolia Yili Industrial Group Co., Ltd. to develop a new method for screening bitter peptides based on machine learning, which used this method to screen and verify the bitter peptides that cause bitterness in milk from ultra-high temperature treatment. The results were published in Food Chemistry.

Screening and validation methods. Photo courtesy of Dalian Chemical Properties

In general, UHT treated milk becomes bitter with the extension of storage time, and protein degradation to produce bitter peptides may be the trigger. In addition to their effects on flavor and nutrition, bitter peptides have also been shown to have antibacterial, anti-inflammatory, antioxidant and other biological functions. Therefore, high-throughput screening of bitter peptides in complex matrices is of great significance for nutritional research and drug development.

In this work, the team built an extended benchmark dataset and a new combination of feature factors, and established a bitter peptide classification prediction model based on the lightweight gradient booster algorithm, and the prediction accuracy of bitter peptide in the independent test set could reach 90.3%. The model was used to predict 180 potential bitter peptides in spoiled UHT treated milk, of which 11 were reported bitter peptides.

This work combines machine learning with omics techniques to propose a complete workflow for the identification and prediction of bitter peptides, revealing the molecular mechanism of bitterness in milk storage during UHT treatment, and at the same time providing a new method for the screening of functional peptides. (Source: Sun Danning, China Science News)

Related paper information:https://doi.org/10.1016/j.foodchem.2023.137288



Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button