5. Mortality, Morbidity, and Pace-of-Aging Clocks Capture Related but Distinct Aspects of Aging

Mortality/morbidity clocks are trained on outcomes (time-to-death or disease incidence) or composite phenotypes linked to those outcomes, rather than solely on chronological age (65, 66). They aim to weigh components of molecular state by their relevance to risk, not just time. Their success in prediction does not by itself establish that the features used by the clock are causal. They are attractive because they approximate what we ultimately care about: what matters most for survival and health. They also expose a limitation of chronological clocks (10, 63): some interventions may modestly shift age-correlated signatures yet strongly affect survival, or vice versa. An additional class of clocks measure a rate of change, revealing how rapidly an organism ages rather than what its current biological age is. Such clocks are also very useful, as they are predictive of aging outcomes. A limitation of some of these clocks is that, by design, they are enriched for late-life biomarkers that may capture pathological processes rather than the underlying aging/deleteriome state. They may also be population-specific, because molecular signatures that predict mortality, morbidity, or pace of aging can reflect prevalent diseases, exposures, and social contexts in the training cohort. Chronological clocks may better predict earlier aging events, upstream of pathology.