Polygenic risk scores are not genetic predispositions
Furrer and Turkheimer’s correspondence in Nature Human Behaviour takes aim at a persistent conceptual error in how PRS are framed by scientists and journalists.
A PRS is a sum of genetic variants weighted by their statistical association with a trait — a risk estimate derived from a population-level model, not an internal property of an individual. Change the reference population, the environmental context, or the statistical model, and the same genome produces a different score. This is fundamentally different from high-penetrance variants like the HTT mutation for Huntington’s disease, where a well-characterized biological pathway links genotype to phenotype.
Calling a PRS a “genetic predisposition” appeals to genetic essentialism — the intuition that genetic information reveals stable, internal essences that define who a person is. Both authors admit to using the term in their own past work, evidence of how reflexive the framing is.
The score is statistical, not dispositional. And the language matters because essentialist framing shapes how people interpret risk, how clinicians counsel patients, and how policy gets made.