Skin DNA Methylation Encodes Multidimensional Facial Aging Phenotypes with Distinct Biological Architectures

ABSTRACT 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.