Integrated flexible DNA methylation-chromatin segmentation modeling enhances epigenomic state annotation

DNA methylation and histone modifications together shape the cell-type-specific epigenomic landscape. To enable a more comprehensive genome-wide annotation, we developed EpiSegMixMeth (ESMM), the first truly integrative segmentation model combining chromatin marks and DNA methylation. ESMM extends hidden Markov models with flexible read count distributions and state duration modeling. Applied to 154 high-quality human epigenomes from the IHEC EpiAtlas, ESMM substantially improves the annotation of broad heterochromatic regions-covering over 60% of the genome, that are frequently missed by chromatin-only models. Additionally, it precisely defines the boundaries of narrow regulatory elements and resolves local chromatin state transitions during cell differentiation. Notably, we demonstrate that DNA methylation can substitute for missing repressive histone marks in segmentation, ensuring robust annotation across diverse cell types. In memory B-cell development, ESMM reveals fine-scale chromatin state shifts that align with 3D genome architecture changes. Our results highlight the power of integrating DNA methylation into genome segmentation and provide a valuable resource for dissecting cell-type-specific epigenomic regulation.