Aging clocks are only reliable when several agree

An aging clock is a mathematical model that estimates biological age from biological data, such as patterns of DNA methylation. These models are trained on large datasets and learn which combinations of biological features correlate with age, disease risk or life expectancy. They are increasingly used to evaluate whether anti-aging therapies have any measurable effect.

Clocks often disagree with each other

The problem: there are now dozens of these clocks, and they frequently disagree. The researchers took previously published studies in which a single clock showed a positive effect of an intervention, then added more clocks to the same analysis. In most cases, those additional clocks did not reach statistical significance. A single positive clock, the findings suggest, is insufficient evidence on its own.

The team then analysed clock behaviour in positive control datasets: situations already known to accelerate aging. From this, they propose three criteria for greater confidence in a result. First, reliable clocks built on principal component analysis respond. Second, clocks trained on mortality risk or pace of aging respond. Third, multiple clocks point in the same direction.

What this means in practice

For anyone tracking longevity therapies, this matters. A supplement or intervention that moves only one aging clock provides limited evidence under this framework. The study, published in the journal Aging Cell, also suggests that some earlier single-clock findings may warrant reexamination. That is not a small caveat: it calls into question the evidence base of multiple published studies. Whether those interventions genuinely work therefore remains uncertain for now.

Read the original article

Search terms to explore further: epigenetic clock intervention study, DNA methylation biological age, aging clock validation