Integrative Proteomic and Metabolomic Signatures of Accelerated PhenoAge in the UK Biobank

Abstract

Aging is accompanied by molecular changes across multiple biological systems that contribute to functional decline and increased disease risk, but the underlying mechanisms and inter-individual variation remain poorly understood. We investigated whether multi-omics integration can reveal coordinated molecular processes associated with accelerated PhenoAge, a clinical biomarker-based estimate of biological aging. Using UK Biobank data from ~20,000 participants, we integrated plasma proteomics and metabolomics with PhenoAge. Individuals were stratified by extreme PhenoAge acceleration (PhenoAgeAccel) to enhance biological contrast. We applied a supervised multi-omics integration framework (DIABLO) to identify correlated molecular features that distinguish accelerated from slower PhenoAge. Proteomics alone provided strong discrimination between aging groups, whereas metabolomics showed weaker performance. Integrating both modalities did not substantially improve classification accuracy but revealed four multi-omics modules: immune-inflammatory, lipid-vascular, nutrient-metabolic, and HDL-apolipoprotein pathways. The dominant signature reflected immune and inflammatory activation with metabolic support, consistent with established aging processes, while additional components highlighted lipid transport, vascular signaling, and nutrient regulation. Nearly half of DIABLO-selected proteins overlapped curated aging and senescence databases, supporting relevance to aging. Together, these results show that multi-omics integration enables the identification of coordinated proteomic-metabolomic axes associated with accelerated aging processes, enhancing biological interpretability beyond single-omics analyses.