Background: Atopic dermatitis (AD) is a chronic inflammatory skin disease characterized by immune dysregulation and epithelial barrier dysfunction. Although transcriptional changes in AD skin are increasingly well characterized, DNA methylation patterns remain less well understood.
Methods: Here, we present an integrated analysis of matched DNA methylomes, transcriptomes, and microbiomes from lesional (n = 40), adjacent non-lesional (n = 38), and healthy control (n = 40) skin, using complementary cell-type-adjusted models.
Results: We identified thousands of differentially methylated regions (DMRs) across all pairwise contrasts (lesional vs. healthy: n = 13,514; lesional vs. non-lesional: n = 4591; non-lesional vs. healthy: n = 1716), including both hyper- and hypo-methylated regions with balanced effect sizes. A core subset of DMRs persisted after methylation-based adjustment, whereas the extensive lesional vs. non-lesional set was largely composition-driven. Integration with transcriptomic co-expression networks linked DMRs to immune-epithelial modules, and 225 DMR-gene pairs showed significant anti-correlation (FDR < 0.05). Lesional skin with dominant Staphylococcus aureus colonization differed at 92 DMRs compared with absent-colonized skin, of which 70 also overlapped with local severity. Pathway analyses consistently highlighted Rho GTPase and actin-junctional programs across analytic layers, suggesting that Rho GTPase signaling is a central integrator of immune, epithelial, and microbial interactions in AD.
Conclusions: Our study underscores the importance of epigenomic remodeling in AD and highlights potential avenues for precision intervention in chronic inflammatory skin disease.
Keywords: atopic dermatitis; methylomics; microbiome dysbiosis; transcriptomics.
© 2026 The Author(s). Allergy published by European Academy of Allergy and Clinical Immunology and John Wiley & Sons Ltd.