A Bayesian Markov generator-matrix model fitted to NHANES (n=23,844) and replicated in the Health and Retirement Study finds that approximately 92% of Gompertz mortality acceleration — the exponential rise in death risk with age — is driven by a latent component undetectable by standard blood biomarker axes. Mendelian randomization across two proteomic platforms (UKB-PPP, deCODE) confirmed known causal proteins (LPA, IL6R) but showed null causal effects for inflammatory, renal, and IGF-axis markers despite their strong predictive associations with mortality. Critically, cellular reprogramming reversed chronological epigenetic clocks by 11–22 years but failed to reverse the causality-enriched DamAge clock.
This preprint, not yet peer-reviewed, represents one of the most methodologically ambitious attempts to formally separate mortality prediction from mortality causation in aging science — a distinction the field has long conflated. The finding that popular epigenetic clocks and plasma biomarkers are largely associative rather than mechanistically upstream of death should temper enthusiasm for longevity interventions targeting these markers as proxies. The reprogramming result is particularly sobering: reversing the epigenetic clock face without reversing damage-linked methylation patterns suggests Yamanaka-factor approaches may be resetting measurable noise rather than core aging drivers. Limitations include reliance on observational cohorts for the mortality model and the inherent assumptions of Mendelian randomization. If replicated and peer-validated, this framework would be genuinely paradigm-shifting for biomarker development and intervention prioritization in geroscience.
