Amino acid-based biological age clock and its implications for human health and aging.

Abstract

Amino acids are fundamental to human physiology, yet their impact on growth, development, and aging remains elusive. Here, we introduce AmiAge, a biological age predictor constructed using a Random Forest model trained on the concentrations of 18 amino acids across individuals aged 1 to 89 years. Leveraging data from 9 studies encompassing over 11,000 in-house and more than 270,000 publicly available samples with diverse demographic and genetic backgrounds, AmiAge demonstrates robust accuracy. The deviation between AmiAge and chronological age (AmiAge Gap) correlates strongly with established aging biomarkers, disease risk, and clinical outcomes. Individuals with higher gaps exhibit increased frailty, telomere attrition, and elevated incidence of age-related diseases. To enhance clinical utility, we distilled AmiAge into an 8-amino acid model (including alanine, glutamine, glycine, histidine, leucine, phenylalanine, tyrosine, and valine). Our findings suggest that this simple, scalable amino acid clock offers a valuable complement to existing biological aging metrics, with potential applications in personalized health management and aging research.