Abstract
The oral cavity harbours one of the most diverse microbial communities in the human body, and disruption of this ecosystem is now known to contribute to caries, periodontal disease, oral cancer and a range of systemic diseases. High-throughput sequencing now characterises these communities in unprecedented detail, but the resulting data are high-dimensional, sparse, compositional and noisy, which constrains classical statistics. Artificial intelligence (AI), specifically machine learning (ML) and deep learning (DL), provides powerful tools to model this complexity, identify microbial signatures and convert community profiles into clinically useful predictions. This review outlines the state of the art of AI-powered oral microbiome analysis. It begins with the biology of the oral microbiome and the sequencing and bioinformatic pipelines used to process samples into feature tables. It then reviews the algorithms most often used for microbiome data, including regularised regression, random forests, convolutional networks and representation-learning networks, and highlights their use in periodontal disease, dental caries, oral cancer, gastrointestinal cancer and the oral–systemic axis. Common issues such as small and heterogeneous cohorts, batch effects, poor external validation and the balance between predictive performance and interpretability are discussed. Finally, future directions including explainable AI, multi-omics integration, standardised reference resources and prospective clinical evaluation are outlined. Driven by a mechanism-aware and rigorously tested approach, the integration of AI and oral microbiome analysis is poised to underpin a future of non-invasive diagnostics and personalized oral care.
Keywords: Artificial Intelligence, Deep Learning, Machine Learning, Metagenomics, Oral Microbiome, Precision Dentistry.