Skin DNA Methylation Encodes Multidimensional Facial Aging Phenotypes with Distinct Biological Architectures
Whether distinct visible aging traits, e.g., wrinkling, pigmentation, and inflammation, reflect shared or independent epigenetic programs remains unknown; existing clocks compress aging into a single chronological axis, leaving the phenotype-specific architecture of cutaneous aging uncharacterized. Here, we integrate AI-derived facial phenotypes with skin DNA methylation profiles from 706 individuals to develop EpiVision, a panel of 21 epigenetic predictors spanning structural, pigmentary, inflammatory, and textural aging traits. Predictors reveal shared and trait-specific pathways, including developmental patterning, epithelial remodeling, hormonal signaling, and UV damage responses, and capture environmentally induced acceleration in sun-exposed skin alongside lifestyle and topical treatment-associated variation. These findings establish that visible skin aging comprises molecularly distinct axes with shared regulatory substrates and trait-specific drivers, providing a scalable epigenetic framework for intervention evaluation and aging biology research.