Powerful three-sample genome-wide design and robust statistical inference in summary-data Mendelian randomization

2019 ◽  
Vol 48 (5) ◽  
pp. 1478-1492 ◽  
Author(s):  
Qingyuan Zhao ◽  
Yang Chen ◽  
Jingshu Wang ◽  
Dylan S Small
Keyword(s):  
Odds Ratio ◽  
Statistical Power ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Density Lipoprotein ◽  
P Value ◽  
Genome Wide ◽  
A Genome ◽  
Summary Data ◽  
Genetic Instruments

Abstract Background Summary-data Mendelian randomization (MR) has become a popular research design to estimate the causal effect of risk exposures. With the sample size of GWAS continuing to increase, it is now possible to use genetic instruments that are only weakly associated with the exposure. Development We propose a three-sample genome-wide design where typically 1000 independent genetic instruments across the whole genome are used. We develop an empirical partially Bayes statistical analysis approach where instruments are weighted according to their strength; thus weak instruments bring less variation to the estimator. The estimator is highly efficient with many weak genetic instruments and is robust to balanced and/or sparse pleiotropy. Application We apply our method to estimate the causal effect of body mass index (BMI) and major blood lipids on cardiovascular disease outcomes, and obtain substantially shorter confidence intervals (CIs). In particular, the estimated causal odds ratio of BMI on ischaemic stroke is 1.19 (95% CI: 1.07–1.32, P-value <0.001); the estimated causal odds ratio of high-density lipoprotein cholesterol (HDL-C) on coronary artery disease (CAD) is 0.78 (95% CI: 0.73–0.84, P-value <0.001). However, the estimated effect of HDL-C attenuates and become statistically non-significant when we only use strong instruments. Conclusions A genome-wide design can greatly improve the statistical power of MR studies. Robust statistical methods may alleviate but not solve the problem of horizontal pleiotropy. Our empirical results suggest that the relationship between HDL-C and CAD is heterogeneous, and it may be too soon to completely dismiss the HDL hypothesis.

scholarly journals Causal Association Between Birth Weight and Atrial Fibrillation: Evidence from a Two-Sample Mendelian Randomization Analysis

2020 ◽  
Author(s):  
Songzan Chen ◽  
Fangkun Yan ◽  
Tian Xu ◽  
Yao Wang ◽  
Kaijie Zhang ◽  
...  
Keyword(s):  
Atrial Fibrillation ◽  
Birth Weight ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Sensitivity Analyses ◽  
P Value ◽  
Causal Association ◽  
Genome Wide ◽  
A Genome ◽  
Mr Study

Abstract Background Although several observational studies have shown an association between birth weight (BW) and atrial fibrillation (AF), controversy remains. In this study, we aimed to explore the role of elevated BW on the etiology of AF. Methods A two-sample Mendelian randomization (MR) study was designed to infer the causality. The genetic data on the associations of single nucleotide polymorphisms (SNPs) with BW and AF were separately obtained from two large-scale genome-wide association study with up to 321,223 and 1,030,836 individuals respectively. SNPs were identified at a genome-wide significant level (p-value < 5 × 10− 8). The inverse variance-weighted (IVW) with fixed effects method was performed to obtain causal estimates as our primary analysis. MR-Egger regression was conducted to assess the pleiotropy and sensitivity analyses with various statistical methods were applied to evaluate the robustness of the results. Results In total, 122 SNPs were identified as the genetic instrumental variables. MR analysis revealed a causal effect of elevated BW on AF (OR = 1.21, 95% CI = 1.13–1.29, p-value = 2.39 × 10− 8). The MR-Egger regression suggested no evidence of directional pleiotropy (intercept = 0.00, p-value = 0.62). All the results in sensitivity analyses were consistent with the primary result, which confirmed the causal association between BW and AF. Conclusions The findings from the two-sample MR study indicate a causal effect of elevated BW on AF. This suggests a convenient and effective method to ease the burden of AF by reducing the number of newborns with elevated BW.

scholarly journals Genetic Architecture of Abdominal Aortic Aneurysm in the Million Veteran Program

Circulation ◽  
2020 ◽  
Vol 142 (17) ◽  
pp. 1633-1646 ◽  
Author(s):  
Derek Klarin ◽  
Shefali Setia Verma ◽  
Renae Judy ◽  
Ozan Dikilitas ◽  
Brooke N. Wolford ◽  
...  
Keyword(s):  
Blood Pressure ◽  
Family History ◽  
Odds Ratio ◽  
Genome Wide Association Study ◽  
Mendelian Randomization ◽  
Risk Variants ◽  
Genome Wide ◽  
Abdominal Aortic ◽  
A Genome ◽  
Current Screening

Background: Abdominal aortic aneurysm (AAA) is an important cause of cardiovascular mortality; however, its genetic determinants remain incompletely defined. In total, 10 previously identified risk loci explain a small fraction of AAA heritability. Methods: We performed a genome-wide association study in the Million Veteran Program testing ≈18 million DNA sequence variants with AAA (7642 cases and 172 172 controls) in veterans of European ancestry with independent replication in up to 4972 cases and 99 858 controls. We then used mendelian randomization to examine the causal effects of blood pressure on AAA. We examined the association of AAA risk variants with aneurysms in the lower extremity, cerebral, and iliac arterial beds, and derived a genome-wide polygenic risk score (PRS) to identify a subset of the population at greater risk for disease. Results: Through a genome-wide association study, we identified 14 novel loci, bringing the total number of known significant AAA loci to 24. In our mendelian randomization analysis, we demonstrate that a genetic increase of 10 mm Hg in diastolic blood pressure (odds ratio, 1.43 [95% CI, 1.24–1.66]; P =1.6×10 −6 ), as opposed to systolic blood pressure (odds ratio, 1.06 [95% CI, 0.97–1.15]; P =0.2), likely has a causal relationship with AAA development. We observed that 19 of 24 AAA risk variants associate with aneurysms in at least 1 other vascular territory. A 29-variant PRS was strongly associated with AAA (odds ratio PRS , 1.26 [95% CI, 1.18–1.36]; P PRS =2.7×10 −11 per SD increase in PRS), independent of family history and smoking risk factors (odds ratio PRS+family history+smoking , 1.24 [95% CI, 1.14–1.35]; P PRS =1.27×10 −6 ). Using this PRS, we identified a subset of the population with AAA prevalence greater than that observed in screening trials informing current guidelines. Conclusions: We identify novel AAA genetic associations with therapeutic implications and identify a subset of the population at significantly increased genetic risk of AAA independent of family history. Our data suggest that extending current screening guidelines to include testing to identify those with high polygenic AAA risk, once the cost of genotyping becomes comparable with that of screening ultrasound, would significantly increase the yield of current screening at reasonable cost.

scholarly journals Causal Inference for Heritable Phenotypic Risk Factors Using Heterogeneous Genetic Instruments

2020 ◽  
Author(s):  
Jingshu Wang ◽  
Qingyuan Zhao ◽  
Jack Bowden ◽  
Gilbran Hemani ◽  
George Davey Smith ◽  
...  
Keyword(s):  
Risk Factors ◽  
Complex Traits ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Association Studies ◽  
Genome Wide Association ◽  
Genome Wide Association Studies ◽  
Genetic Associations ◽  
Genome Wide ◽  
Genetic Instruments

Over a decade of genome-wide association studies have led to the finding that significant genetic associations tend to spread across the genome for complex traits. The extreme polygenicity where "all genes affect every complex trait" complicates Mendelian Randomization studies, where natural genetic variations are used as instruments to infer the causal effect of heritable risk factors. We reexamine the assumptions of existing Mendelian Randomization methods and show how they need to be clarified to allow for pervasive horizontal pleiotropy and heterogeneous effect sizes. We propose a comprehensive framework GRAPPLE (Genome-wide mR Analysis under Pervasive PLEiotropy) to analyze the causal effect of target risk factors with heterogeneous genetic instruments and identify possible pleiotropic patterns from data. By using summary statistics from genome-wide association studies, GRAPPLE can efficiently use both strong and weak genetic instruments, detect the existence of multiple pleiotropic pathways, adjust for confounding risk factors, and determine the causal direction. With GRAPPLE, we analyze the effect of blood lipids, body mass index, and systolic blood pressure on 25 disease outcomes, gaining new information on their causal relationships and the potential pleiotropic pathways.

scholarly journals Bipolar disorder and cannabis use: a bidirectional two-sample Mendelian randomization study

2020 ◽  
Author(s):  
Oskar Hougaard Jefsen ◽  
Maria Speed ◽  
Doug Speed ◽  
Søren Dinesen Østergaard
Keyword(s):  
Bipolar Disorder ◽  
Standard Deviation ◽  
Odds Ratio ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Cannabis Use ◽  
Genome Wide ◽  
Log Odds ◽  
The Relationship ◽  
Mr Study

AbstractAimsCannabis use is associated with a number of psychiatric disorders, however the causal nature of these associations has been difficult to establish. Mendelian randomization (MR) offers a way to infer causality between exposures with known genetic predictors (genome-wide significant single nucleotide polymorphisms (SNPs)) and outcomes of interest. MR has previously been applied to investigate the relationship between lifetime cannabis use (having ever used cannabis) and schizophrenia, depression, and attention-deficit / hyperactivity disorder (ADHD), but not bipolar disorder, representing a gap in the literature.MethodsWe conducted a two-sample bidirectional MR study on the relationship between bipolar disorder and lifetime cannabis use. Genetic instruments (SNPs) were obtained from the summary statistics of recent large genome-wide association studies (GWAS). We conducted a two-sample bidirectional MR study on the relationship between bipolar disorder and lifetime cannabis use, using inverse-variance weighted regression, weighted median regression and Egger regression.ResultsGenetic liability to bipolar disorder was significantly associated with an increased risk of lifetime cannabis use: scaled log-odds ratio (standard deviation) = 0.0174 (0.039); P-value = 0.00001. Genetic liability to lifetime cannabis use showed no association with the risk of bipolar disorder: scaled log-odds ratio (standard deviation) = 0.168 (0.180); P-value = 0.351. The sensitivity analyses showed no evidence for pleiotropic effects.ConclusionsThe present study finds evidence for a causal effect of liability to bipolar disorder on the risk of using cannabis at least once. No evidence was found for a causal effect of liability to cannabis use on the risk of bipolar disorder. These findings add important new knowledge to the understanding of the complex relationship between cannabis use and psychiatric disorders.

scholarly journals The putative causal effect of type 2 diabetes in risk of cataract: a Mendelian randomization study in East Asian

2021 ◽  
Author(s):  
Haoyang Zhang ◽  
Xuehao Xiu ◽  
Angli Xue ◽  
Yuedong Yang ◽  
Yuanhao Yang ◽  
...  
Keyword(s):  
Type 2 Diabetes ◽  
Genetic Correlation ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Association Studies ◽  
Population Based ◽  
P Value ◽  
Genome Wide Association Studies ◽  
Genome Wide

AbstractBackgroundThe epidemiological association between type 2 diabetes and cataract has been well-established. However, it remains unclear whether the two diseases share a genetic basis, and if so, whether this reflects a causal relationship.MethodsWe utilized East Asian population-based genome-wide association studies (GWAS) summary statistics of type 2 diabetes (Ncase=36,614, Ncontrol=155,150) and cataract (Ncase=24,622, Ncontrol=187,831) to comprehensively investigate the shared genetics between the two diseases. We performed 1. linkage disequilibrium score regression (LDSC) and heritability estimation from summary statistics (ρ-HESS) to estimate the genetic correlation and local genetic correlation between type 2 diabetes and cataract; 2. multiple Mendelian randomization (MR) analyses to infer the putative causality between type 2 diabetes and cataract; and 3. Summary-data-based Mendelian randomization (SMR) to identify candidate risk genes underling the causality.ResultsWe observed a strong genetic correlation (rg=0.58; p-value=5.60×10−6) between type 2 diabetes and cataract. Both ρ-HESS and multiple MR methods consistently showed a putative causal effect of type 2 diabetes on cataract, with estimated liability-scale MR odds ratios (ORs) at around 1.10 (95% confidence interval [CI] ranging from 1.06 to 1.17). In contrast, no evidence supports a causal effect of cataract on type 2 diabetes. SMR analysis identified two novel genes MIR4453HG (βSMR=−0.34, p-value=6.41×10−8) and KCNK17 (βSMR=−0.07, p-value=2.49×10−10), whose expression levels were likely involved in the putative causality of type 2 diabetes on cataract.ConclusionsOur results provided robust evidence supporting a causal effect of type 2 diabetes on the risk of cataract in East Asians, and posed new paths on guiding prevention and early-stage diagnosis of cataract in type 2 diabetes patients.Key MessagesWe utilized genome-wide association studies of type 2 diabetes and cataract in a large Japanese population-based cohort and find a strong genetic overlap underlying the two diseases.We performed multiple Mendelian randomization models and consistently disclosed a putative causal effect of type 2 diabetes on the development of cataract.We revealed two candidate genes MIR4453HG and KCNK17 whose expression levelss are likely relevant to the causality between type 2 diabetes and cataract.Our study provided theoretical fundament at the genetic level for improving early diagnosis, prevention and treatment of cataract in type 2 diabetes patients in clinical practice

scholarly journals Extending causality tests with genetic instruments: an integration of Mendelian Randomization and the Classical Twin Design

2017 ◽  
Author(s):  
Camelia C. Minică ◽  
Conor V. Dolan ◽  
Dorret I. Boomsma ◽  
Eco de Geus ◽  
Michael C. Neale
Keyword(s):  
Statistical Power ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Twin Design ◽  
Twin Model ◽  
Polygenic Scores ◽  
Exposure Variable ◽  
Causality Tests ◽  
Instrument Bias ◽  
Genetic Instruments

ABSTRACTMendelian Randomization (MR) is an important approach to modelling causality in non-experimental settings. MR uses genetic instruments to test causal relationships between exposures and outcomes of interest. Individual genetic variants have small effects, and so, when used as instruments, render MR liable to weak instrument bias. Polygenic scores have the advantage of larger effects, but may be characterized by direct pleiotropy, which violates a central assumption of MR.We developed the MR-DoC twin model by integrating MR with the Direction of Causation twin model. This model allows us to test pleiotropy directly. We considered the issue of parameter identification, and given identification, we conducted extensive power calculations. MR-DoC allows one to test causal hypotheses and to obtain unbiased estimates of the causal effect given pleiotropic instruments (polygenic scores), while controlling for genetic and environmental influences common to the outcome and exposure. Furthermore, MR-DoC in twins has appreciably greater statistical power than a standard MR analysis applied to singletons, if the unshared environmental effects on the exposure and the outcome are uncorrelated. Generally, power increases with: 1) decreasing residual exposure-outcome correlation, and 2) decreasing heritability of the exposure variable.MR-DoC allows one to employ strong instrumental variables (polygenic scores, possibly pleiotropic), guarding against weak instrument bias and increasing the power to detect causal effects. Our approach will enhance and extend MR’s range of applications, and increase the value of the large cohorts collected at twin registries as they correctly detect causation and estimate effect sizes even in the presence of pleiotropy.

scholarly journals The Effect of Circulating Zinc, Selenium, Copper and Vitamin K1 on COVID-19 Outcomes: A Mendelian Randomization Study

Nutrients ◽  
2022 ◽  
Vol 14 (2) ◽  
pp. 233
Author(s):  
Maria K. Sobczyk ◽  
Tom R. Gaunt
Keyword(s):  
Critical Illness ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Sensitivity Analyses ◽  
European Ancestry ◽  
P Value ◽  
Vitamin K1 ◽  
Inverse Variance ◽  
A Genome

Background & Aims: Previous results from observational, interventional studies and in vitro experiments suggest that certain micronutrients possess anti-viral and immunomodulatory activities. In particular, it has been hypothesized that zinc, selenium, copper and vitamin K1 have strong potential for prophylaxis and treatment of COVID-19. We aimed to test whether genetically predicted Zn, Se, Cu or vitamin K1 levels have a causal effect on COVID-19 related outcomes, including risk of infection, hospitalization and critical illness. Methods: We employed a two-sample Mendelian Randomization (MR) analysis. Our genetic variants derived from European-ancestry GWAS reflected circulating levels of Zn, Cu, Se in red blood cells as well as Se and vitamin K1 in serum/plasma. For the COVID-19 outcome GWAS, we used infection, hospitalization or critical illness. Our inverse-variance weighted (IVW) MR analysis was complemented by sensitivity analyses including a more liberal selection of variants at a genome-wide sub-significant threshold, MR-Egger and weighted median/mode tests. Results: Circulating micronutrient levels show limited evidence of association with COVID-19 infection, with the odds ratio [OR] ranging from 0.97 (95% CI: 0.87–1.08, p-value = 0.55) for zinc to 1.07 (95% CI: 1.00–1.14, p-value = 0.06)—i.e., no beneficial effect for copper was observed per 1 SD increase in exposure. Similarly minimal evidence was obtained for the hospitalization and critical illness outcomes with OR from 0.98 (95% CI: 0.87–1.09, p-value = 0.66) for vitamin K1 to 1.07 (95% CI: 0.88–1.29, p-value = 0.49) for copper, and from 0.93 (95% CI: 0.72–1.19, p-value = 0.55) for vitamin K1 to 1.21 (95% CI: 0.79–1.86, p-value = 0.39) for zinc, respectively. Conclusions: This study does not provide evidence that supplementation with zinc, selenium, copper or vitamin K1 can prevent SARS-CoV-2 infection, critical illness or hospitalization for COVID-19.

scholarly journals Circulating Vitamin E Levels and Risk of Coronary Artery Disease and Myocardial Infarction: A Mendelian Randomization Study

Nutrients ◽  
2019 ◽  
Vol 11 (9) ◽  
pp. 2153 ◽  
Author(s):  
Tao Wang ◽  
Lin Xu
Keyword(s):  
Myocardial Infarction ◽  
Coronary Artery Disease ◽  
Coronary Artery ◽  
Vitamin E ◽  
Mendelian Randomization ◽  
Density Lipoprotein ◽  
Lipoprotein Cholesterol ◽  
Genome Wide ◽  
A Genome ◽  
Artery Disease

Observational studies have reported a cardioprotective effect of vitamin E whereas intervention trials failed to confirm its beneficial effects, and even some reported adverse effects of vitamin E supplements on coronary artery disease (CAD). To clarify, we conducted a two-sample mendelian randomization study to investigate causal association of vitamin E with the risk of CAD. Three single nucleotide polymorphisms (SNPs) identified in a genome-wide analysis study including 7781 individuals of European descent, rs964184, rs2108622, and rs11057830 were used as the genetic instruments for vitamin E. Data for CAD/myocardial infarction (MI) were available from Coronary ARtery DIsease Genome wide Replication and Meta-analysis (CARDIoGRAM) plus The Coronary Artery Disease (C4D) Genetics consortium. The effect of each SNP on CAD/myocardial infarction (MI) was weighted by its effect on serum vitamin E (mg/L), and results were pooled to give a summary estimates for the effect of increased vitamin E on risk of CAD/MI. Based on 3 SNPs each 1 mg/L increase in vitamin E was significantly associated with CAD (odds ratio (OR) 1.05, 95% confidence interval (CI) 1.03–1.06), MI (OR 1.04, 95% CI 1.03–1.05), elevated low-density lipoprotein cholesterol (0.021 standard deviations (SD), 95% CI 0.016, 0.027), triglycerides (0.026 SD, 95% CI 0.021, 0.031), and total cholesterol (0.043 SD, 95% CI 0.038, 0.048) and lower levels of high-density lipoprotein cholesterol (−0.019 SD 95% CI −0.024, −0.014). Our findings indicate that higher vitamin E may increase the risk of CAD/MI and the safety and efficacy of vitamin E supplementation use should be reevaluated.

scholarly journals Causally Associations of Blood Lipids Levels with COVID-19 Risk: Mendelian Randomization Study

2020 ◽  
Author(s):  
Kun Zhang ◽  
Yan Guo ◽  
Zhuo-Xin Wang ◽  
Jing-Miao Ding ◽  
Shi Yao ◽  
...  
Keyword(s):  
Blood Lipids ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Low Density Lipoprotein ◽  
Density Lipoprotein ◽  
Lipoprotein Cholesterol ◽  
Gene Set Analysis ◽  
P Value ◽  
Comorbid Condition ◽  
Gene Set

AbstractBackgroundCoronavirus disease 2019 (COVID-19) is a global pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2). It has been found that coronary artery disease (CAD) is a comorbid condition for COVID-19. As the risk factors of CAD, whether blood lipids levels are causally related to increasing susceptibility and severity of COVID-19 is still unknown.DesignWe performed two-sample Mendelian Randomization (MR) analyses to explore whether dyslipidemia, low density lipoprotein cholesterol (LDL-c), high density lipoprotein cholesterol (HDL-c), triglyceride (TG) and total cholesterol (TC) were causally related to COVID-19 risk and severity. The GWAS summary data of blood lipids involving in 188,578 individuals and dyslipidemia in a total of 53,991 individuals were used as exposures, respectively. Two COVID-19 GWASs including 1,221 infected patients and 1,610 severe patients defined as respiratory failure were employed as outcomes. Based on the MR estimates, we further carried out gene-based and gene-set analysis to explain the potential mechanism for causal effect.ResultsThe MR results showed that dyslipidemia was casually associated with the susceptibility of COVID-19 and induced 27% higher odds for COVID-19 infection (MR-IVW OR = 1.27, 95% CI: 1.08 to 1.49, p-value = 3.18 × 10−3). Moreover, the increasing level of blood TC will raise 14 % higher odds for the susceptibility of COVID-19 (MR-IVW OR = 1.14, 95% CI: 1.04 to 1.25, p-value = 5.07 × 10−3). Gene-based analysis identified that ABO gene was associated with TC and the gene-set analysis found that immune processes were involved in the risk effect of TC.ConclusionsWe obtained three conclusions: 1) Dyslipidemia is casually associated with the susceptibility of COVID-19; 2) TC is a risk factor for the susceptibility of COVID-19; 3) The different susceptibility of COVID-19 in specific blood group may be partly explained by the TC concentration in diverse ABO blood groups.

scholarly journals No Causal Effect of Telomere Length on Ischemic Stroke and Its Subtypes: A Mendelian Randomization Study

Cells ◽  
2019 ◽  
Vol 8 (2) ◽  
pp. 159 ◽  
Author(s):  
Weijie Cao ◽  
Xingang Li ◽  
Xiaoyu Zhang ◽  
Jie Zhang ◽  
Qi Sun ◽  
...  
Keyword(s):  
Ischemic Stroke ◽  
Telomere Length ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Genome Wide Association Studies ◽  
Large Artery ◽  
Genome Wide ◽  
Inverse Variance ◽  
A Genome ◽  
Weighted Median

Background: Epidemiological studies observing inconsistent associations of telomere length (TL) with ischemic stroke (IS) are susceptible to bias according to reverse causation and residual confounding. We aimed to assess the causal association between TL, IS, and the subtypes of IS, including large artery stroke (LAS), small vessel stroke (SVS), and cardioembolic stroke (CES) by performing a series of two-sample Mendelian randomization (MR) approaches. Methods: Seven single nucleotide polymorphisms (SNPs) were involved as candidate instrumental variables (IVs), summarized from a genome-wide meta-analysis including 37,684 participants of European descent. We analyzed the largest ever genome-wide association studies of stroke in Europe from the MEGASTROKE collaboration with 40,585 stroke cases and 406,111 controls. The weighted median (WM), the penalized weighted median (PWM), the inverse variance weighted (IVW), the penalized inverse variance weighted (PIVW), the robust inverse variance weighted (RIVW), and the Mendelian randomization-Egger (MR-Egger) methods were conducted for the MR analysis to estimate a causal effect and detect the directional pleiotropy. Results: No significant association between genetically determined TL with overall IS, LAS, or CES were found (all p > 0.05). SVS was associated with TL by the RIVW method (odds ratio (OR) = 0.72, 95% confidence interval (CI): 0.54–0.97, p = 0.028), after excluding rs9420907, rs10936599, and rs2736100. Conclusions: By a series of causal inference approaches using SNPs as IVs, no strong evidence to support the causal effect of shorter TL on IS and its subtypes were found.

Sign in / Sign up

Export Citation Format

Share Document