Ancestry Paradox in Polygenic Scores

Updated: 2026.01.12 16D ago 7 sources
Polygenic scores trained on European datasets underperform in non‑European populations, yet institutions often deny biologically meaningful group differences. Embryo‑selection tools thus work best for Europeans, creating a two‑tier system while exposing a contradiction between practice and prevailing narratives. — It forces regulators, clinicians, and media to confront ancestry‑specific performance and its ethical and political implications for equity and how we talk about race and genetics.

Sources

Ten Myths About Human Genetics That Refuse to Die
Davide Piffer 2026.01.12 90% relevant
The article explicitly disputes simplistic readings of ancestry plots and the overclaiming of polygenic scores across populations; that directly maps to the existing concern that PGS trained on Europeans perform unevenly across ancestries and are often misused in public discourse.
Can we detect polygenic selection within Europe without being fooled by population structure?
Davide Piffer 2026.01.07 82% relevant
Piffer’s discussion of stratification confounding and Yuan et al.’s explicit modeling of population structure connects to the existing idea that PGS performance and interpretation vary by ancestry and that claims of recent selection can be artifacts of training data and structure.
Genetic evidence for race differences in behaviour
Davide Piffer 2026.01.04 60% relevant
The bear study underscores how population‑specific evolutionary histories create distinct genomic signatures tied to behaviour; that dynamic maps onto the existing concern that polygenic predictors and their interpretation are ancestry‑dependent and that naive cross‑population inference risks error or misuse.
Why Africans Can Look Closer to the Human–Chimp Ancestor Under Some Metrics
Davide Piffer 2026.01.02 85% relevant
The post describes how metric and ascertainment choices yield ancestry‑specific appearances (Africans looking 'more ancestral')—the same phenomenon that produces polygenic‑score performance differences across ancestries and the ethical/policy tensions discussed in the existing idea.
Let That Skin In: Ancient DNA and the Evolution of Human Skin Colour
Davide Piffer 2025.12.01 87% relevant
Both items grapple with polygenic‑score portability and ancestry dependence: this article attempts to validate PGS on ancient genomes using a multi‑ancestry Pan‑UKBB pigmentation GWAS and by recovering known geographic clines, directly engaging the same methodological concerns about applying PGS across ancestries and time.
How Embryo Selection Technology exposes the Transferability Paradox
2025.10.07 100% relevant
The article’s 'Race Transferability Paradox' framing and Herasight’s own benchmarks showing significantly lower PGS accuracy outside white British ancestry.
Race: a social destruction of a biological concept | Biology & Philosophy
2010.01.12 65% relevant
Sesardic argues that biological taxonomy of human groups resists simplistic eliminativist critiques; that bears on how we interpret polygenic score performance differences across ancestries (the 'ancestry paradox') because conceptual clarity about population structure informs whether observed differences are treated as biologically meaningful or as artefacts of study design.
← Back to All Ideas