Artificial Intelligence-Driven Multi-Omics Diagnostic Pipelines for Infectious, Neurodegenerative, and Metabolic Diseases: From Biomarker Discovery to Precision Medicine and Digital Twin Healthcare

Abstract

The combination of AI and multi-omics has ushered in a new, revolutionary era in disease diagnosis and precision medicine. This review aims to summarize the current status of the artificial intelligence (AI)-based multi-omics diagnostic workflows employed in three disease paradigms: infectious, neurodegenerative, and metabolic diseases, and critically evaluate their translational potential from the discovery of biomarkers to digital twin healthcare systems. Machine Learning (ML) and Deep Learning (DL) algorithms, as well as Explainable Artificial Intelligence (XAI) algorithms, are explored for their application in linking genomics, transcriptomics, proteomics, metabolomics, and microbiomics datasets to enable high-dimensional molecular phenotyping. Harmonisation strategies for multi-omics data, AI driven feature selection, molecular pathway elucidation using graph neural networks (GNNs) and transformer architectures, and the development of digital twin models for personalised, dynamic health simulation are among the key themes. We delve deeper into the regulatory, ethical, and equity issues arising from the deployment of AI-omics systems across heterogeneous clinical settings. The review ends with a strategy for integrating validated AI-omics pipelines into the next-generation precision therapeutics and global health infrastructure.