Blog post dissects a Science paper arguing lifespan heritability rises from ~25-35% to ~50% only by modeling a hypothetical world with zero extrinsic mortality.
Key Takeaways
Traditional twin studies peg lifespan heritability at 23-35% (Danish and Swedish cohorts); the Science paper’s 46-57% figure applies only to a simulated accident/murder/disease-free world.
The paper’s core move: build a parametric death-probability model, fit it to historical data, then re-run simulations with extrinsic mortality dialed to zero and measure heritability on that synthetic data.
Heritability rising when environmental noise is removed is mathematically trivial; the paper’s value is in quantifying how much it rises and offering a method to project heritability for modern cohorts with lower baseline extrinsic mortality.
The author flags that Science’s editorial format strips out math and systematic methodology, making the paper’s actual procedures hard to verify without extensive appendix archaeology.
Fraternal twin simulation relies on correlating parameters at 0.5 to reflect 50% shared DNA, but the paper is vague on the exact sampling procedure, leaving a reproducibility gap.
