Exploration of Novel Antimicrobial Agents against Foodborne Pathogens via a Deep Learning Approach

The emergence of antibiotic-resistant bacteria poses a severe threat to food safety and human health, necessitating an urgent search for novel antimicrobial agents that can be applied in the food industry. This study utilizes a deep learning approach to establish the optimal models for antibacterial activity against foodborne pathogens, particularly Escherichia coli and Staphylococcus aureus, as well as for predicting carcinogenicity. These optimal models are applied to screen natural products from the COCONUT database, resulting in the identification of 130 compounds with both antibacterial activity and noncarcinogenic properties. Two natural products, bis(hexamethylene)triamine and N-phenethylbiguanide, are selected for experimental validation of their antibacterial activity. The confirmation of antimicrobial properties validates the reliability of the models developed in this study. By providing an innovative approach for identifying antimicrobial agents for foodborne pathogens, this research offers new insights for discovering effective antimicrobials in an efficient manner.