Researchers at King’s College London have developed AI-driven metabolomic aging clocks that leverage blood metabolite data to forecast health and lifespan. Their study, published in Science Advances, involved 17 machine learning algorithms tested on data from over 225,000 UK Biobank participants aged 40 to 69. The most effective algorithm was found to be Cubist regression, as it accurately predicted biological age, defined as “MileAge,” by analyzing metabolites—small molecules from metabolic processes.
The study revealed that individuals with an older metabolite-predicted age than their chronological age faced higher health risks, including frailty, chronic illnesses, and increased mortality. Accelerated aging correlated with shorter telomeres, cellular aging markers linked to age-related diseases. Conversely, younger metabolite-predicted ages showed only weak associations with good health.
Dr. Julian Mutz, the study’s lead author, emphasized the potential of these aging clocks to inform preventative health strategies and assist individuals in making lifestyle adjustments for better health outcomes. The study underscores the promise of big data analytics in enhancing biological aging assessments that could ultimately guide health decisions and interventions.
