PheWAS-based clustering of Mendelian Randomisation instruments reveals distinct mechanism-specific causal effects between obesity and educational attainment

Abstract

Mendelian Randomisation (MR) is a statistical method that estimates causal effects between risk factors and common complex diseases using genetic instruments. Heritable confounders, pleiotropy and heterogeneous causal effects violate MR assumptions and can lead to biases. To tackle these, we propose an approach employing a PheWAS-based clustering of the MR instruments (PWC-MR). We apply this method to revisit the surprisingly large causal effect of body mass index (BMI) on educational attainment (EDU): â = -0.19 [-0.22, -0.16]. As a first step of PWC-MR, we clustered 324 BMI-associated genetic instruments based on their association profile across 407 traits in the UK Biobank, which yielded six distinct groups. The subsequent cluster-specific MR revealed heterogeneous causal effect estimates on EDU. A cluster strongly enriched for traits related to socio-economic position yielded the largest BMI-on-EDU causal effect estimate (â = -0.49 [-0.56, -0.42]) whereas a cluster enriched for primary impact on body-mass had the smallest estimate (â = -0.09 [-0.13, -0.05]). Several follow-up analyses confirmed these findings: (i) within-sibling MR results (â = -0.05 [-0.09, -0.01]); (ii) MR for childhood BMI on EDU (â = -0.03 [-0.06, -0.002]); (iii) step-wise multivariable MR (MVMR) (â = -0.06 [-0.09, -0.04]) where time spent watching television and past tobacco smoking (two proxies for potential confounders) were jointly modelled. Through a detailed examination of the BMI-EDU causal relationship we demonstrated the utility of our PWC-MR approach in revealing distinct pleiotropic pathways and confounder mechanisms.

Distinct pathway-based effects of blood pressure and body mass index on cardiovascular traits: comparison of novel Mendelian Randomization approaches

“…Similarly to Darrous et al (2023) 9 , we conclude that our new method offers complementary, orthogonal SNP stratification to the existing colocalisation approach as evidenced by very weak overlap. Comparison of MR results using both approaches suggests agreement for the dominant role of vasculature-related SNPs in determining the left ventricular stroke volume and the risk of type 2 diabetes, but discordance with regards to coronary heart disease, and influence of different BMI pathways on atrial fibrillation.…”

“…Of chief interest in our study, causal effects vary due to endpoint phenotypes being de facto composites, representing divergent underlying biological mechanisms covered by different genetic instruments, which unlike bias can improve understanding of disease aetiology and help design better targeted interventions. Three broad types of approaches have been used so far when studying biological sources of heterogeneity: direct clustering based on SNP associations with exposure and outcome 5,6 , clustering of variant associations across a set of traits [7][8][9] , or instrument clustering informed by tissue gene expression patterns [10][11][12][13] . In particular, a biological hypothesis-driven approach proposed by Leyden et al (2022) 11 clusters genetic instruments for body mass index (BMI) based on the tissue (brain or adipose) where a given BMI SNP is found to colocalise with an expression quantitative trait locus (eQTL).…”

Testing the causal relationships of physical activity and sedentary behaviour with mental health and substance use disorders: a Mendelian randomisation study

“…Notably, these traits could represent alternative pathways through which genetic variants linked to physical activity may affect mental health and could therefore be possible sources of horizontal pleiotropy. Future research could further explore the role of pleiotropic effects using multivariable MR to test the direct effects of physical activity on mental health and substance use disorders after controlling for potential confounding factors (e.g., intelligence, educational attainment, body mass index), as well as novel MR approaches such as PheWAS-based clustering of Mendelian Randomisation instruments [57]. Fourth, we found evidence of a bidirectional causal relationship between sedentary behaviour and schizophrenia.…”

MR-link-2: pleiotropy robustcisMendelian randomization validated in four independent gold-standard datasets of causality

“…As it is usually impossible to ensure that a genetic variant only acts through the chosen exposure (7). MR methods are constantly being developed to ensure MR estimates are robust (8)(9)(10)(11)(12)(13). Generally, these methods use tens to hundreds of independent locations on the genome in a meta-analysis to mitigate violations of the relevance assumption and exclusion restriction with the hope that independent instruments would lead to independent biases which cancel each other out.…”