Analysis of published anti-aging intervention studies found that single-clock epigenetic results frequently fail to replicate when additional clock models are applied, with no further clocks reaching statistical significance. The authors propose three criteria for credible results: response in reliable principal-component-based clocks, response in mortality-trained or pace-of-aging clocks, and concordant direction across multiple clocks. For anyone using methylation age as an endpoint, this reframes a single favorable clock reading as insufficient evidence of biological benefit.
Key Points
- Single-clock intervention findings rarely replicate across additional clock models
- Reliable PC-based and mortality-trained clocks should respond together
- Concordant direction across multiple clocks distinguishes true effects from noise
Longevity Analysis
Methylation-based age estimates have become a default outcome measure in both clinical trials and consumer testing, yet clock-specific biology and measurement noise produce the same apparent result as a genuine intervention effect. Reading these outputs correctly means treating a favorable number from one algorithm as a signal requiring corroboration rather than a verdict on regeneration or mortality risk. Interventions worth sustaining are those that move several reliable clocks in the same direction, and that standard protects against building long-term protocols on statistical artifacts.
Original published by Wiley Aging Cell, by Daniel S. Borrus, Raghav Sehgal, Jenel Fraij Armstrong, Jessica Kasamoto, John Gonzalez, Albert Higgins‐Chen .