Combined polygenic scores boost IQ prediction

Updated: 2026.05.10 9D ago 16 sources
Aggregating GWAS results for intelligence and related traits (notably years of education) produces multipolygenic scores that explain substantially more variation in measured intelligence than single‑trait scores — the review reports combined scores explaining over 10% of variance and accounting for ~20% of the heritable component. This quantitative jump transforms polygenic scores from weak correlates into variables of practical predictive use in longitudinal and policy research. — Greater predictive power makes polygenic intelligence scores relevant to education policy, clinical uses, reproductive decisions, and debates over fairness and privacy.

Sources

Culture Links, 5/10/2026
Arnold Kling 2026.05.10 90% relevant
Razib Khan summarizes a new Nature paper from Reich's lab claiming recent selection in Europeans for cognitive/education traits and discusses polygenic-score performance — this directly connects to the broader idea that polygenic scores are improving prediction of cognitive outcomes and are entering public debate.
Embryo Selection and Frontier Genomics with Dr. Alex Young – Manifold #111
Steve Hsu 2026.05.07 74% relevant
The podcast addresses recent gains in predictive power (family‑based methods, imputation across relatives) that mirror the trend of combining datasets and methods to strengthen polygenic prediction for cognitive and educational outcomes.
Did Harsh Seasons Make Complex Societies?
Davide Piffer 2026.05.04 80% relevant
This article engages the same empirical tool — polygenic scores for educational attainment/cognitive proxies — but finds they do not mediate the link from harsh seasons to archaeological civilization stage; that nuance directly qualifies the expectation that stronger PGS prediction implies causal genetic pathways to complex societies.
Is Gusev right about Family GWAS? Signal-maxxing using cross-population LD
Davide Piffer 2026.04.23 82% relevant
This article interrogates how family‑GWAS (direct‑effect) summary statistics — which show much weaker signals when clumped at genome‑wide thresholds — might nonetheless be combined or reweighted using cross‑population linkage‑disequilibrium (LD) structure to recover adaptation/selection signals; that directly connects to the broader idea that combining or reprocessing polygenic signals can materially change predictive power for cognitive traits (actor: Guan et al family GWAS; method: cross‑population LD 'signal‑maxxing').
Podcast: The Akbari–Piffer controversy
2026.04.22 80% relevant
Discussion on polygenic versus single‑gene selection, confounds, and methodology directly ties to the idea that combining polygenic predictors changes effect sizes and interpretability for cognitive traits — a technical point at the heart of the controversy.
Top 10 Most Replicable Findings from Behavior Genetics, Part 1: Findings #1-5
Steve Stewart-Williams 2026.04.18 70% relevant
The article reiterates the strong, replicable genetic contribution to psychological traits (e.g., ~50% heritability for intelligence), which is the empirical background that makes polygenic prediction possible and politically salient; it connects the Plomin synthesis to the broader trend of improving genomic prediction referenced by that idea.
Recent Human Evolution in Europe
Steve Stewart-Williams 2026.04.17 90% relevant
The paper directly reports coordinated, population‑level shifts in combinations of alleles that today predict complex traits including cognitive performance (polygenic scores). That empirical claim strengthens the existing idea that polygenic scores are becoming powerful and societally consequential tools (Akbari et al., 15,836 West Eurasian samples; provided selection coefficients and PGS trends).
Ancient DNA reveals pervasive directional selection across West Eurasia
Isegoria 2026.04.16 85% relevant
The paper reports directional selection on combinations of alleles that today predict cognitive performance, which links directly to the claim that combined polygenic scores meaningfully predict cognitive traits; the ancient‑DNA evidence shows those combinations have been shifting in frequency over millennia (15,836 samples, selection coefficients estimated at 9.7M variants), reinforcing the idea that polygenic scores capture biologically real, historically‑active variation.
Educational Attainment Polygenic Scores Track Civilization Stage, Not Just Chronology
Davide Piffer 2026.04.09 62% relevant
By working with educational‑attainment polygenic scores in ancient samples, the article relates to the broader finding that combined polygenic indices improve cognitive prediction; it extends that empirical project back into prehistory and asks what those scores mean in different social contexts.
Interpreting Polygenic Prediction of Cognitive Ability
Tyler Cowen 2026.04.07 90% relevant
The article summarizes a paper that constructs and validates a within‑family polygenic score (PGS) for general cognitive ability showing high within‑family retention (attenuation ≈ 0.88) and a corrected within‑family association of about 0.45 with latent general ability, directly supporting the existing idea that combined PGS materially improve prediction of cognitive traits and life outcomes.
Polygenic Score Prediction Within and Between Sibling Pairs for Intelligence, Cognitive Abilities, and Educational Traits From Childhood to Early Adulthood | Published in Intelligence & Cognitive Abilities
2026.04.04 75% relevant
The study reports robust PGS prediction across development and compares cognitive traits to anthropometrics, reinforcing that aggregated/combined PGSs can strongly predict cognitive outcomes in population samples even as within-family prediction is smaller, which connects to claims that scaling GWAS/PGS increases predictive power.
The new genetics of intelligence - PMC
2026.04.04 92% relevant
The article reports that recent large GWAS have identified many variants whose aggregated effects (polygenic scores) account for a substantial share of measured cognitive variance — precisely the empirical basis summarized by the existing idea that combined polygenic scores substantially improve prediction of IQ and related life outcomes.
Associations between common genetic variants and income provide insights about the socio-economic health gradient | Nature Human Behaviour
2025.01.28 82% relevant
This paper extends the polygenic‑score approach to income: the authors build an 'Income Factor' polygenic index (N=668,288 Europeans, 162 loci) that explains ~1–5% of income variance — directly paralleling claims that combined polygenic scores can predict complex social outcomes and thus connects to debates about genetic prediction of socioeconomic traits.
Genetic variation, brain, and intelligence differences | Molecular Psychiatry
2021.02.02 90% relevant
The article reviews DNA‑based heritability estimates and the discovery of loci from large genome‑wide association studies (GWAS), directly supporting the existing claim that combining GWAS results into polygenic scores improves prediction of cognitive test performance; the authors report modest but growing predictive power and discuss genetic correlations with education and health outcomes.
Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence - PubMed
2018.07.04 90% relevant
The paper (Savage et al., 2018) reports new genome‑wide significant loci and functional links from a meta‑analysis of 269,867 people, directly increasing the variant set available to build polygenic scores for cognitive ability and therefore improving predictive power — precisely the mechanism captured by the existing idea about combined polygenic scores boosting IQ prediction.
The new genetics of intelligence | Nature Reviews Genetics
2018.01.08 100% relevant
Key claim in the review: “More than 10% of the variance in intelligence can be predicted by multipolygenic scores derived from GWAS of both intelligence and years of education.”
← Back to all ideas