The Translation Gap in Longevity Genomics

Recent million-person GWAS meta-analyses demonstrate that longevity is highly polygenic, with polygenic risk scores (PRS) predicting up to a five-year lifespan variance 1. However, translating these PRS into actionable interventions has stalled. The primary limitation is that current scoring methods mathematically conflate two distinct mechanistic axes: a metabolic-vascular axis that drives disease avoidance (e.g., APOE, LPA), and a neuro-immune/DNA repair axis that promotes intrinsic cellular resilience (e.g., FOXO3, DNA repair pathways) 2. Aggregating disease-avoidance and active resilience into a single score obscures the mechanistic heterogeneity required for therapeutic targeting.

Hypothesis

I propose the generation of a Decoupled Resilience Polygenic Score (drPRS), which explicitly regresses out loci associated with cardiovascular, metabolic, and oncogenic disease risk, isolating variants exclusively linked to lifespan extension (including rare centenarian variants in IGF-1/Wnt pathways 3).

My core hypothesis is that high drPRS variants function mechanistically as "epigenetic stress capacitors." Unlike disease-avoidance loci, which minimize primary cellular damage, drPRS loci actively buffer the epigenome against stress-induced degradation. Consequently, I hypothesize that drPRS will inversely correlate with the pace of biological aging—measured via dynamic epigenetic clocks like DunedinPACE 4—exclusively under conditions of high environmental or metabolic stress, demonstrating a profound gene-environment (GxE) interaction.

Mechanistic Reasoning

While traditional PRS models assume additive genetic effects on baseline mortality risk 5, resilience networks likely operate as dynamic, stress-responsive feedback loops. Similar to how engineered synthetic gene oscillators actively prevent cellular degeneration in yeast by toggling state-specific maintenance programs 6, endogenous resilience pathways in humans (e.g., FOXO3-mediated autophagy, base excision repair) may remain relatively dormant under low-stress conditions.

Under low environmental stress, basal homeostatic mechanisms are sufficient, and the epigenetic clock ticks at a baseline rate regardless of one's drPRS. However, under high stress (e.g., systemic inflammation, high cumulative smoking pack-years, elevated LDL), high drPRS individuals possess the transcriptomic elasticity required to prevent stress-induced epigenetic drift. They do not avoid the metabolic insult; they prevent the insult from permanently accelerating their DunedinPACE.

Proposed Methodology and Falsification

To test this:

1. Model Construction: Utilize Bayesian priors from known mortality risk factors to penalize and remove metabolic/vascular trait loci from a massive longevity GWAS dataset, utilizing both common SNPs and rare exomic variants.
2. Cohort Stratification: Apply this drPRS to a longitudinal dataset with sequential biospecimen age data (e.g., UK Biobank or ALSPAC) 7. Stratify the cohort into low, medium, and high metabolic/environmental stress profiles.
3. Evaluation: Measure the acceleration of DunedinPACE across a 5-10 year follow-up.

Falsification criteria:

• If the drPRS uniformly slows DunedinPACE across all individuals regardless of environmental stress, the "epigenetic capacitor" (GxE) hypothesis is false; the variants merely dictate a slower basal rate of aging.
• If the drPRS fails to statistically decouple from cardiometabolic disease incidence, it implies that "resilience" is an artifact of unmeasured disease avoidance, nullifying the existence of a distinct maintenance axis.

By decoupling these axes, we move computational genomics away from black-box correlation and closer to mechanism-based longevity therapeutics.