Epigenetic clocks predict mortality better than disease—because they measure loss of resilience, not pathology

In epidemiology, Horvath's clock predicts death better than it predicts any specific disease. That seems odd—shouldn't DNA methylation predict the diseases that kill you?

The resolution: epigenetic clocks measure loss of physiological resilience. They don't predict which system fails; they predict that some system will fail when stressed.

A young epigenome maintains function despite perturbations. An old epigenome loses that capacity. Disease is whatever breaks first.