Abstract
Methylomics has emerged as a central framework for understanding gene regulation in development and disease, yet the rapid expansion of profiling technologies, computational integration methods, and clinical applications has outpaced comprehensive synthesis. This review addresses that gap by systematically examining current advances across the full methylomics pipeline, from data generation to clinical translation. We draw on evidence from large-scale consortium datasets and benchmarking studies of multi-omics integration methods including MOFA, DIABLO, and deep learning architectures, single-cell and spatial methylomic technologies, long-read sequencing platforms (Oxford Nanopore, PacBio HiFi), and cell-free DNA (cfDNA) liquid biopsy approaches. The review further surveys methylation dysregulation across major disease domains, including cancer, cardiovascular disease, neurological disorders, and autoimmune conditions. Integrating methylomic data with transcriptomic and chromatin accessibility layers, particularly in spatial and single-cell contexts, substantially improves the resolution of disease-associated regulatory mechanisms. cfDNA methylation profiling emerges as a cross-disease, non-invasive monitoring platform with broad diagnostic potential, supported by machine learning-based deconvolution. We conclude that while technological barriers are diminishing, standardization of analytical workflows, population diversity in reference datasets, and regulatory alignment remain the principal challenges for translating methylomics advances into broadly accessible precision medicine.