Keywords
How to Cite
Abstract
Although much progress has been made on the detection, treatment, and prevention of TB, it remains one of the most lethal infectious diseases globally. M. tuberculosis remains a complex pathogen with highly variable host–pathogen interactions, which have contributed to its persistence over time. These interactions influence susceptibility, immune response, disease course, and treatment outcome, and can vary widely between individuals.
While traditional studies have provided useful mechanistic insights, studying individual genes, proteins, or pathways may not be sufficient to understand the complex and heterogeneous disease process. New multi-omics platforms facilitate the study of various biological systems in parallel. Genomics, epigenomics, transcriptomics, proteomics, and metabolomics have all undergone tremendous advances recently, and further advances in single-cell and spatial multi-omics are expected to reveal new dimensions of pathogen biology, host immune response, cellular heterogeneity, and TB-specific signatures.
However, interpreting these large and heterogeneous datasets requires specialized tools that allow the extraction of biologically relevant and clinically actionable knowledge from the data. AI and machine learning have emerged as powerful tools for multi-omics data integration, biomarker discovery, risk stratification, and individualized diagnostic and therapeutic applications.
Recent progress in host–pathogen interactions, multi-omics, and AI offer unprecedented opportunities to develop precise approaches for TB control. There remain, however, several obstacles to address before these advances can be effectively translated to clinical practice. These include the challenge of understanding the underlying biology, inadequate sample sizes and cohort representativeness, methodological inconsistencies, difficulties in data integration and model interpretability, and unequal availability of technological infrastructure.
In this review, we explore the potential of multi-omics and AI to better understand the pathophysiology of TB and its host–pathogen interactions, and support biomarker discovery, risk prediction, and patient-specific disease management. We further consider the main scientific, technical, and implementation barriers that need to be overcome to advance towards clinical application.