Participation bias in the UK Biobank distorts genetic associations and downstream analyses

Schoeler, Tabea; Speed, Doug; Porcu, Eleonora; Pirastu, Nicola; Pingault, Jean‐Baptiste; Kutalik, Zoltán

doi:10.1038/s41562-023-01579-9

Cited by 111 publications

(65 citation statements)

References 46 publications

(58 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, one potential limitation of this approach is that it does not produce SNP effect sizes on a similar scale to case-control studies, which creates challenges in the interpretation and some applications requiring effect sizes. While weighted least squares is a common approach to account for non-random study participation (32), it did not give promising results in our analysis. Excluding individuals who do not know about parental health from the analysis and residualizing GWAX on genetic associations with parental health awareness both reduced the spurious AD-EA genetic correlation.…”

Section: Discussionmentioning

confidence: 72%

“…The Jansen approach yielded a significant but positive genetic correlation with coronary artery disease (cor = 0.14, p = 1.5E-6). To reduce participation bias, we followed Schoeler et al (32) and conducted a weighted GWAS on parental AD status. We trained a LASSO regression model on whether a survey participant reported parental illnesses using a random subset of UKB samples and then performed weighted GWAS on the remaining samples (Methods).…”

Section: Reducing Biases In Gwaxmentioning

confidence: 99%

“…We used a weighted GWAS approach to account for non-random survey participation. Following Schoeler et al (32), we used 14 variables to train a participation prediction model. These variables comprise five continuous ones: age, body mass index, weight, height, and the age at which full-time education was completed, and nine categorical variables: household size (1-7 or more individuals), sex (male or female), alcohol consumption frequency (never to daily), smoking habits (never, previous, or current smoker), employment status (employed, economically inactive, retired, or unemployed), income brackets (from <18k to >100k), obesity classification (underweight, healthy weight, overweight, or obese), general health status (poor, fair, or good), and urbanization level (from village/hamlet to urban).…”

Section: Approaches For Bias Reduction In Gwaxmentioning

confidence: 99%

See 2 more Smart Citations

Pervasive biases in proxy GWAS based on parental history of Alzheimer’s disease

Wu,

Sun,

Zheng

et al. 2023

Preprint

View full text Add to dashboard Cite

Almost every recent Alzheimer's disease (AD) genome-wide association study (GWAS) has performed meta-analysis to combine studies with clinical diagnosis of AD with studies that use proxy phenotypes based on parental disease history. Here, we report major limitations in current GWAS-by-proxy (GWAX) practices due to uncorrected survival bias and non-random participation of parental illness survey, which cause substantial discrepancies between AD GWAS and GWAX results. We demonstrate that current AD GWAX provide highly misleading genetic correlations between AD risk and higher education which subsequently affects a variety of genetic epidemiologic applications involving AD and cognition. Our study sheds important light on the design and analysis of mid-aged biobank cohorts and underscores the need for caution when interpreting genetic association results based on proxy-reported parental disease history.

show abstract

Section: Discussionmentioning

confidence: 72%

Section: Reducing Biases In Gwaxmentioning

confidence: 99%

Section: Approaches For Bias Reduction In Gwaxmentioning

confidence: 99%

See 1 more Smart Citation

Pervasive biases in proxy GWAS based on parental history of Alzheimer’s disease

Wu,

Sun,

Zheng

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Third, information bias could be present due to undiagnosed or unreported cases of schizophrenia and other self-reported health-related covariates. Fourth, potential participation bias should be acknowledged, as our sample was slightly biased towards individuals with higher socio-economic status, similar to other genetic studies (Schoeler et al, 2023). Extrapolation to younger populations should be made cautiously.…”

Section: Strengths and Weaknessesmentioning

confidence: 81%

Schizophrenia polygenic risk score and type 2 diabetes onset in older adults with no schizophrenia diagnosis

Shamsutdinova

Ajnakina

Roberts

et al. 2023

Psychiatric Genetics

View full text Add to dashboard Cite

Objectives An association between type 2 diabetes (T2DM) and schizophrenia has long been observed, and recent research revealed presence of shared genetic factors. However, epidemiological evidence was inconsistent, some reported insignificant contribution of genetic factors to T2DM-schizophrenia comorbidity. Prior works studied people with schizophrenia, particularly, antipsychotic-naive patients, or those during the first psychotic experience to limit schizophrenia-related environmental factors. In contrast, we controlled such factors by utilizing a general population sample of individuals undiagnosed with schizophrenia. We hypothesized that if schizophrenia genetics impact T2DM development and such impact is not fully mediated by schizophrenia-related environment, people with high polygenic schizophrenia risk would exhibit elevated T2DM incidence. Methods Using a population-representative sample of adults aged ≥50 from English Longitudinal Study of Ageing (n = 5968, 493 T2DM cases, average follow-up 8.7 years), we investigated if schizophrenia polygenic risk score (PGS-SZ) is associated with T2DM onset. A proportional hazards model with interval censoring was adjusted for age and sex (Model 1), and age, sex, BMI, hypertension, cardiovascular diseases, exercise, smoking, depressive symptoms and T2DM polygenic risk score (Model 2). According to the power calculations, hazard rates > 1.14 per standard deviation in PGS-SZ could be detected. Results We did not observe a significant association between PGS-SZ and T2DM incidence (hazard ratio 1.04; 95% CI 0.93–1.15; and 1.01, 95% CI 0.94–1.09). Conclusion Our results suggest low contribution of the intrinsic biological mechanisms driven by the polygenic risk of schizophrenia on future T2DM onset. Further research is needed.

show abstract

“…However, this manuscript is not without limitations. First, this is a cross-sectional study in a European ancestry subsample of individuals whose caregivers volunteered to participate in the research study (Schoeler et al, 2023). This limits study generalizability, which may be especially important for better understanding potential relationships between facets of academic achievement and factors related to health and well being.…”

Section: Discussionmentioning

confidence: 99%

Associations Between Polygenic Scores for Cognitive and Non-cognitive Factors of Educational Attainment and Measures of Behavior, Psychopathology, and Neuroimaging in the Adolescent Brain Cognitive Development Study

Gorelik,

Paul,

Miller

et al. 2023

Preprint

View full text Add to dashboard Cite

BackgroundBoth cognitive and non-cognitive (e.g., traits like curiosity) factors are critical for social and emotional functioning and independently predict educational attainment. These factors are heritable and genetically correlated with a range of health-relevant traits and behaviors in adulthood (e.g., risk-taking, psychopathology). However, whether these associations are present during adolescence, and to what extent these relationships diverge, could have implications for adolescent health and well-being.MethodsUsing data from 5,517 youth of European ancestry from the ongoing Adolescent Brain Cognitive DevelopmentSMStudy, we examined associations between polygenic scores (PGS) for cognitive and non-cognitive factors and outcomes related to cognition, socioeconomic status, risk tolerance and decision-making, substance initiation, psychopathology, and brain structure.ResultsCognitive and non-cognitive PGSs were both positively associated with cognitive performance and family income, and negatively associated with ADHD and severity of psychotic-like experiences. The cognitive PGS was also associated with greater risk-taking, delayed discounting, and anorexia, as well as lower likelihood of nicotine initiation. The cognitive PGS was further associated with cognition scores and anorexia inwithin-siblinganalyses, suggesting these results do not solely reflect the effects of assortative mating or passive gene-environment correlations. The cognitive PGS showed significantly stronger associations with cortical volumes than the non-cognitive PGS and was associated with right hemisphere caudal anterior cingulate and pars-orbitalis inwithin-siblinganalyses, while the non-cognitive PGS showed stronger associations with white matter fractional anisotropy and a significantwithin-siblingassociation for right superior corticostriate-frontal cortex.ConclusionsOur findings suggest that PGSs for cognitive and non-cognitive factors show similar associations with cognition and socioeconomic status as well as other psychosocial outcomes, but distinct associations with regional neural phenotypes in this adolescent sample.

show abstract

Participation bias in the UK Biobank distorts genetic associations and downstream analyses

Cited by 111 publications

References 46 publications

Pervasive biases in proxy GWAS based on parental history of Alzheimer’s disease

Pervasive biases in proxy GWAS based on parental history of Alzheimer’s disease

Schizophrenia polygenic risk score and type 2 diabetes onset in older adults with no schizophrenia diagnosis

Associations Between Polygenic Scores for Cognitive and Non-cognitive Factors of Educational Attainment and Measures of Behavior, Psychopathology, and Neuroimaging in the Adolescent Brain Cognitive Development Study

Contact Info

Product

Resources

About