Current technologies, such as smart watches, apps and fitness trackers, already give some indirect measurements of health, and these indicators may reflect likely healthspan. These assist interventions to some degree, helping people adhere to the only activities known to optimise ageing: exercise, diet and sleep regimens. However, much more granular information is required if meaningful increases to healthspan are to be achieved.
New methods — blood7 or urine tests8, and saliva and stool-sourced microbiome analysis — can purportedly reflect age-related variations in blood markers, measuring objective indicators and revealing differential “ageotypes”, and potentially warnings of premature ageing.9 These biomarkers are not yet validated in humans but when they are, they will provide measures of “biological age” - a more useful indicator of how long a person can expect to remain in good health than number of years alive. We now know that certain genes can slow ageing in centenarians and lead to healthier old age than their age in years would suggest.10 Furthermore, the body’s different tissues can age at vastly different rates due to genetic or environmental factors.1112
The goal is a biomarker of age whose manipulation restores good health, the way blood pressure does for heart disease. Many candidates exist including transcriptomics, metabolomics, proteomics and DNA methylation. Omics analysis of existing studies of gerotherapies — including rapamycin, metformin and senolytics — is now underway with a view to finding consequent changes in age-related biomarkers. The recent discovery of three senolytics using machine learning algorithms trained solely on published data is just one way in which AI is helping the search for new biomarkers.13 Machine learning will be used to search hundreds of datasets and previous trials to identify other useful factors and patterns.14 Population-level work is aiming to expose the relationships between these clocks and disease risk.15
