Applications of Metagenomics and Artificial Intelligence in Characterizing Antimicrobial Resistance in Livestock: A Systematic Review

Abstract

Antimicrobial resistance (AMR) is an urgent global health threat, intensified by the widespread use of antimicrobials in livestock production. This study synthesizes the current landscape of combining metagenomic sequencing with artificial intelligence (machine learning and deep learning) to characterize, surveil, and predict AMR within the One Health framework. A comprehensive multi-database literature search was conducted, and, following PRISMA guidelines, 10 peer-reviewed studies meeting the inclusion criteria were selected for full synthesis. Metagenomic shotgun sequencing significantly surpasses conventional culture-based methods by directly capturing antimicrobial resistance genes (ARGs) from complex biological communities. AI algorithms substantially outperform traditional bioinformatic tools, achieving high predictive accuracy (AUC-ROC > 0.90) and revealing consistent ARG transfer pathways that link livestock, human, and environmental compartments. Integrating metagenomics with AI delivers a paradigm shift for proactive AMR surveillance. However, standardization, interpretability, and technological adaptation to resource-limited settings-especially in sub-Saharan Africa-remain urgent priorities to inform effective public health policy.