Microbiome-based disease prediction with artificial intelligence: A systematic review

The human microbiome plays a critical role in both physical and mental health, with alterations in its composition and function linked to a wide range of diseases, including neurological, cardiovascular, respiratory, and inflammatory disorders. The rapid growth of microbiome datasets has driven the application of artificial intelligence (AI) for disease diagnostics and data-driven analysis. This review critically examines AI-based approaches used to investigate associations between the microbiome and chronic diseases, covering both classical machine learning and deep learning methods applied to diverse microbiome profiling data. The integration of microbiome data with AI techniques has shown promising performance in identifying disease-specific signatures and improving diagnostic accuracy. However, significant challenges remain, particularly methodological heterogeneity and the lack of standardized analytical pipelines. Notably, 16S ribosomal RNA (16S rRNA) sequencing remains a widely used approach in microbiome research, accounting for 37% of the studies included in this review. A comparative analysis of methodological approaches was also conducted to identify research gaps and evaluate the most effective strategies for microbiome-based disease prediction.