Umeå University's logo

 DNA methylation is an epigenetic mechanism crucial for gene regulation and cell differentiation. The process occurs at ’CG’ motives along the DNA sequence, called CpG sites. Methylation patterns shift during development and in response to diseases, stress, and aging. Their relation to aging is exploited by statistical models known as Epigenetic Clocks. The clocks track the methylation levels of 100 − 102 specific sites and aim to measure an individual’s biological age, rather than chronological age. However, the mechanism driving DNA methylation changes over time, and its effects on organisms remain unknown. To address the gap, we study two CpG density-dependent models: a mean-field model for CpG clusters and a molecular model with single-site resolution. The mean-field model, composed of Ordinary Differential Equations, uses Bayesian inference to fit parameters to experimental DNA methylation data. The molecular model simulates specific genome segments, tracking the methylation state of individual CpG sites. We tested the models on CpG sites associated with EpigeneticClocks. We found that the mean-field model performs well on small clusters (<60CpG sites), while the molecular model reflects known methylation characteristics but overestimates demethylation in real sequences. The results lay the groundwork for improving the models and conducting mechanistic analyses. In the future, we aim to mimic molecular stresses, evaluate gene expression changes related to diseases, and identify vulnerable methylation states and sensitive parameters.