What does an epigenetic clock actually measure?

Small chemical tags on DNA, called methyl groups, attach themselves to thousands of fixed locations in the genome. Over the years these tags shift in a predictable way. An epigenetic clock is an algorithm that reads those patterns and calculates an age figure from them. The first well-known clock, developed by researcher Steve Horvath, used 353 of these sites, trained on more than 8,000 tissue samples from 51 different tissue types.

The number the clock produces is called the 'biological age' by researchers. The difference from your actual calendar age is called age acceleration: a positive value means your cells look biologically older than they are in calendar terms, a negative value means the opposite. That the clock measures something meaningful is shown, among other things, by the fact that reprogrammed stem cells -- adult cells that have in effect been reset to an early, unspecialised state -- score close to zero on the clock. And cancerous tissue scores on average 36 years older than healthy tissue from the same person.

People with an epigenetic age higher than their calendar age have, in observational research, an increased risk of death, cardiovascular disease, cancer and diabetes. Their blood cells also show more signs of inflammatory activity and less active processes for DNA repair and cellular energy production. This provides a biological fingerprint of what 'ageing faster' means at the cellular level. These are, however, associations from observational research; whether the clock also drives ageing or merely signals it has not been proven.

Newer clocks, such as PhenoAge and GrimAge, were not trained on calendar age but on clinical measures of health and mortality. They therefore predict more accurately than older versions who will become ill or die sooner. Yet their value as a biomarker has not been fully confirmed in long-running studies either.

A technical caveat: blood consists of multiple cell types, each with its own methylation pattern. If that composition changes due to age or disease, it affects the clock reading in a way that has nothing to do with 'real' ageing. Researchers are actively working on methods to correct for this. And although clocks are increasingly being used to measure whether lifestyle changes or medications lower biological age, those studies are still small and short-term. Whether a lower clock reading actually means you will live longer or in better health has not yet been demonstrated.