Potential of exudate metabolites to predict microbial populations in beef

This study explores the feasibility of using beef exudate metabolites as non-destructive predictors of total viable count (TVC) in beef stored under vacuumed and refrigerated conditions (4 °C). A total of 55 metabolites were identified, among which 36 metabolites exhibited strong correlations with TVC, with correlation coefficients exceeding |0.5|. Notably, metabolites such as nicotinamide adenine dinucleotide (NAD+), citrate, and hypoxanthine showed the highest correlations with TVC. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment and predictive modeling using Least Absolute Shrinkage and Selection Operator (LASSO) regression demonstrated the effectiveness of grouping metabolites by pathways. The combination of pathways involving phenylalanine, tyrosine, and tryptophan biosynthesis; glycine, serine, and threonine metabolism; along with starch and sucrose metabolism achieved the highest predictive accuracy for TVC (R² = 0.957, RMSEP = 0.298, and RPD = 5.260). A total of 12 metabolites-betaine, choline, creatine, dimethyl sulfone, glucose, glucose-1-phosphate, glycine, phenylalanine, pyruvate, sarcosine, serine, and tyrosine-were used as predictors. These pathways were interconnected through glycolysis/gluconeogenesis, highlighting their role in bacterial energy metabolism and their relevance to early freshness changes. This study highlights the utility of metabolomics as a non-invasive approach to evaluate beef freshness, offering a foundation for improving quality control methods within the meat industry. Further validation under different storage conditions is recommended to broaden its practical applications.