EXPLORING BACTERIAL INTERACTIONS IN FOOD METAGENOMICS: ARTIFICIAL INTELLIGENCE APPROACH

Abstract

Food spoilage poses a major challenge to food security, public health, and economic stability. This study examines microbial interactions in ready-to-eat ham using 16S rRNA metabarcoding and association rule mining (ARM) with machine learning. By analyzing microbial co-occurrence patterns through key metrics—support, confidence, lift, and conviction—this research identifies microbial relationships influencing spoilage dynamics. Notably, Escherichia coli and Klebsiella were strongly associated with spoilage risks, while Lactobacillus and Enterococcus exhibited potential for mitigating spoilage effects. Higher-order associations involving Bacillus, Staphylococcus, and Enterococcus revealed complex microbial networks shaping spoilage processes. These insights enhance the understanding of microbial ecology in food systems, highlighting both spoilage risks and the role of beneficial microbes. This study demonstrates the effectiveness of ARM in metagenomic analysis, uncovering microbial interactions that inform predictive tools for microbial monitoring. By identifying key microbial relationships, it supports spoilage prevention strategies such as inhibiting pathogens or enhancing beneficial microbes. The findings contribute to food safety, quality management, and waste reduction, promoting sustainable food systems through data-driven approaches.