Gene–environment correlations across geographic regions affect genome-wide association studies

Summary

This study demonstrates that gene–environment correlations arising from the geographic clustering of both genetic variants and socioeconomic status meaningfully bias genome-wide association study results for complex traits. Using up to 254,387 British individuals and 43,516 sibling pairs, the authors show that controlling for geographic region substantially reduces heritability estimates for socioeconomic traits and alters genetic correlations across most phenotypes studied. The findings suggest that both passive (birthplace) and active (current address) gene–environment correlations introduce systematic bias into GWAS analyses.

UK applicability

These findings are directly applicable to UK research using biobank data and UK population studies, as the analysis was conducted entirely in British cohorts. The results highlight the importance of geographic adjustment in UK GWAS studies and suggest that previous estimates of genetic contribution to socioeconomic and health-related traits in British populations may have been inflated.

Key measures

Heritability estimates before and after geographic region adjustment; genetic correlations with educational attainment and income; polygenic scores for educational attainment; geographic clustering of genotypes and socioeconomic status

Outcomes reported

The study quantified how geographic region controls affect heritability estimates and genetic correlations in genome-wide association studies across 56 complex traits. It demonstrated that controlling for geographic regions significantly reduced estimated heritability for socioeconomic status-related traits and altered genetic correlations, particularly for body mass index, body fat, sedentary behaviour and substance use.

Topic tags