Refining the generation, interpretation and application of multi-organ, multi-omics biological aging clocks

Abstract

Multi-organ biological aging clocks derived from clinical phenotypes and neuroimaging data have emerged as valuable tools for studying human aging and disease. Plasma proteomics provides an additional molecular dimension to enrich these clocks. In this study, I developed 11 multi-organ proteome-based biological age gaps (ProtBAGs) using 2,448 plasma proteins from 43,498 participants in the UK Biobank. Here I highlight methodological and clinical considerations for developing and using these clocks, including correction for age bias, organ specificity of proteins, sample size and underlying pathologies in the training data, which can affect model generalizability and clinical interpretability. In addition, I integrated 11 ProtBAGs with previously developed nine multi-organ phenotype-based biological age gaps to investigate genetic overlap and causal associations with disease endpoints. Finally, I show that incorporating features across organs improves predictions for systemic disease categories and all-cause mortality. These analyses provide methodological and clinical insights for developing and interpreting these clocks and highlight future avenues toward a multi-organ, multi-omics biological aging clock framework. In developing multi-organ-based and multi-omics-based biological aging clocks and linking them to underlying genetics and clinical phenotypes, Wen demonstrates the value of methodological and clinical considerations for enhancing model generalizability and clinical interpretability.