DNA Methylation as a Programmable Information Layer: From Molecular Marks to Disease State Engineering

DNA methylation has long been regarded as a stable, maintenance-based epigenetic marker. However, this classical binary model struggles to fully explain the dynamic and situational dependence of methylation regulation at the multi-biological level. This review defines DNA methylation as a programmable information layer that systematically integrates the latest advances in three interrelated dimensions of molecular coding, disease status indication, and epigenomic engineering. At the molecular level, this paper describes how the chemical diversity of cytosine modification, the writing-erasing enzyme network, and the three-dimensional structure of chromatin jointly construct a methylated polymorphic coding system and evaluates the performance of emerging sequencing technologies in DNA integrity, reading length, modification resolution, and analytical complexity through a multidimensional scoring framework. At the cellular and clinical levels, this paper comprehensively demonstrates methylation as a quantifiable indicator of cell identity, biological aging and disease status, covering circulating free DNA biomarkers and spatial heterogeneity analysis. Critically, this paper evaluates how the clustered regularly interspaced short palindromic repeats (CRISPR)-based epigenome editing platform achieves causal inference and promotes the transformation of methylation from related biomarkers to functional therapeutic targets. At the same time, persistent challenges such as off-target specificity, in vivo delivery, and spatiotemporal regulation encountered in epigenetic gene editing are discussed. This review reveals the paradigm shift of DNA methylation from passive observation markers to actively engineered regulatory parameters, which has direct therapeutic application prospects.