New AmiAge clock, validated across more than 280,000 samples, correlates biological age gap with frailty, telomere loss, and disease incidence

A newly published study introduces AmiAge, a biological age predictor built on the concentrations of circulating amino acids in the bloodstream. Developed using a machine learning model trained on data spanning individuals aged one to 89 years, AmiAge offers a straightforward and scalable approach to estimating how fast a person is aging at the cellular and physiological level — a question that conventional chronological age alone cannot answer.

The research, published in Nature Communications, draws on nine separate studies comprising more than 11,000 in-house samples and over 270,000 publicly available records representing diverse demographic and genetic backgrounds. The breadth of the dataset lends the model an unusual degree of statistical robustness for a tool in this category.

At its core, AmiAge quantifies the difference between a person’s predicted biological age and their actual chronological age — a figure the researchers call the AmiAge Gap. This gap proved to be a meaningful signal. Individuals with a higher gap, indicating that their bodies appear older than their birth year would suggest, showed greater levels of frailty, accelerated telomere shortening, and a higher incidence of age-related diseases. The AmiAge Gap also correlated strongly with established aging biomarkers and measurable clinical outcomes, lending weight to its validity as a health assessment tool.

To make the tool viable in clinical and research settings, the team refined the model down to eight key amino acids: alanine, glutamine, glycine, histidine, leucine, phenylalanine, tyrosine, and valine. This condensed panel preserves the predictive accuracy of the fuller model while reducing the analytical burden, making AmiAge practical for routine laboratory testing without requiring specialist infrastructure.

Biological age clocks have proliferated in recent years as researchers apply machine learning techniques to ever-wider categories of physiological data, from DNA methylation patterns to protein levels and inflammatory markers. The challenge facing the field is less about producing new clocks and more about demonstrating that any given clock genuinely captures meaningful aspects of the aging process and can reliably track the effects of interventions intended to slow or reverse it. AmiAge addresses this challenge in part by linking its core measure — the gap between biological and chronological age — directly to concrete health outcomes rather than relying solely on internal statistical validation.

The researchers describe AmiAge as a complement to existing biological aging metrics rather than a replacement for them. Its potential applications span personalised health management, epidemiological research, and the assessment of longevity-focused interventions. Because standard amino acid panels are already a routine component of metabolic blood work in many clinical contexts, the tool could be integrated into existing workflows with minimal additional cost or disruption.

Amino acids are fundamental to virtually every biological process, from protein synthesis and cellular repair to immune function and metabolic regulation. Despite this centrality, their relationship to the rate of aging has remained poorly understood. AmiAge represents a structured effort to close that gap, offering researchers and clinicians a new lens through which the biology of aging can be observed, measured, and potentially modified.

Source:  AmiAge: a biological age predictor based on circulating amino acid levels. Nature Communications, 2026. DOI: 10.1038/s41467-026-73371-y