A Primer in Mendelian Randomization Methodology with a Focus on Utilizing Published Summary Association Data

Causal Associations of DNA Methylation and Cardiovascular Disease: A Two-Sample Mendelian Randomization Study

Background

There is growing evidence that concentrations of DNA methylation are associated with cardiovascular disease; however, it is unclear whether this association reflects a causal relationship.

Methods

We utilized a two-sample Mendelian randomization (MR) approach to investigate whether DNA methylation can affect the risk of developing cardiovascular disease in human life. We primarily performed the inverse variance weighted (IVW) method to analyze the causal effect of DNA methylation on multiple cardiovascular diseases. Additionally, to ensure the robustness of our findings, we conducted several sensitivity analyses using alternative methodologies. These analysis methods included maximum likelihood, MR-Egger regression, weighted median method, and weighted model methods.

Results

Inverse variance weighted estimates suggested that an SD increase in DNA methylation Hannum age acceleration exposure increased the risk of cardiac arrhythmias (OR = 1.03, 95% CI 1.00-1.05, p = 0.0290) and atrial fibrillation (OR = 1.03, 95% CI 1.00-1.05, p = 0.0022). We also found that an SD increase in DNA methylation PhenoAge acceleration exposure increased the risk of heart failure (OR = 1.01, 95% CI 1.00-1.03, p = 0.0362). Exposure to DNA methylation-estimated granulocyte proportions was found to increase the risk of hypertension (OR = 1.00, 95% CI 1.00-1.0001, p = 0.0291). Exposure to DNA methylation-estimated plasminogen activator inhibitor-1 levels was found to increase the risk of heart failure (OR = 1.00, 95% CI 1.00-1.00, p = 0.0215).

Conclusion

This study reveals a causal relationship between DNA methylation and CVD. Exposed to high levels of DNA methylation Hannum age acceleration inhabitants with an increased risk of cardiac arrhythmias and atrial fibrillation. DNA methylation PhenoAge acceleration levels exposure levels were positively associated with the increased risk of developing heart failure. This has important implications for the prevention of cardiovascular diseases.

Tissue-specific genetic variation suggests distinct molecular pathways between body shape phenotypes and colorectal cancer

It remains unknown whether adiposity subtypes are differentially associated with colorectal cancer (CRC). To move beyond single-trait anthropometric indicators, we derived four multi-trait body shape phenotypes reflecting adiposity subtypes from principal components analysis on body mass index, height, weight, waist-to-hip ratio, and waist and hip circumference. A generally obese (PC1) and a tall, centrally obese (PC3) body shape were both positively associated with CRC risk in observational analyses in 329,828 UK Biobank participants (3728 cases). In genome-wide association studies in 460,198 UK Biobank participants, we identified 3414 genetic variants across four body shapes and Mendelian randomization analyses confirmed positive associations of PC1 and PC3 with CRC risk (52,775 cases/45,940 controls from GECCO/CORECT/CCFR). Brain tissue-specific genetic instruments, mapped to PC1 through enrichment analysis, were responsible for the relationship between PC1 and CRC, while the relationship between PC3 and CRC was predominantly driven by adipose tissue-specific genetic instruments. This study suggests distinct putative causal pathways between adiposity subtypes and CRC.

Inflammatory cytokines and oral lichen planus: a Mendelian randomization study

Background

Inflammatory cytokines have long been considered closely related to the development of oral lichen planus (OLP), and we further explored the causal relationship between the two by Mendelian randomization (MR) method.

Methods

We performed bidirectional MR analyses by large genome-wide association studies (GWAS). The data included a large-scale OLP dataset, as well as datasets of 41 inflammatory cytokines. All data were obtained from the University of Bristol database, which includes 41 inflammatory cytokines, and the GWAS Catalog database, which includes 91 inflammatory cytokines. OLP data were obtained from the Finngen database, which includes 6411 cases and 405770 healthy controls. We used the inverse variance weighted (IVW) method, MR-Egger method, weighted median method, simple mode method and weighted mode method to analyze the causal relationship between inflammatory cytokines and OLP, and we also combined with sensitivity analysis to further verify the robustness of the results. We performed a meta-analysis of positive or potentially positive results for the same genes to confirm the reliability of the final results.

Results

We primarily used the IVW analysis method, corrected using the Benjamin Hochberg (BH) method. When p<0.00038 (0.05/132), the results are significantly causal; when 0.00038

Conclusion

There is a causal association between OLP and some inflammatory cytokines, which may play an important role in the pathogenesis of OLP and require further attention.

Collapse

Key Words

• Mendelian randomization
• immunity
• inflammation
• inflammatory cytokines
• oral lichen planus

Collapse

MESH Headings

• Humans
• Cytokines/genetics
• Genome-Wide Association Study
• Lichen Planus, Oral/genetics
• Mendelian Randomization Analysis
• Reproducibility of Results

Collapse

Grants Collapse

Affiliation(s)

Xin Chen Department of Oral and Maxillofacial Surgery, Jiangyin People's Hospital Affiliated to Nantong University, Wuxi, Jiangsu, China
Simin Zhang Key Laboratory of Shaanxi Province for Craniofacial Precision Medicine Research, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China Department of General Dentistry, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China
Xiao Wu Endodontic Department, School of Stomatology, Nanjing Medical University, Nanjing, Jiangsu, China
Yuxi Lei Key Laboratory of Shaanxi Province for Craniofacial Precision Medicine Research, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China Department of Emergency Room, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China
Bing Lei Key Laboratory of Shaanxi Province for Craniofacial Precision Medicine Research, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China Department of General Dentistry, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China
Zhibai Zhao Key Laboratory of Shaanxi Province for Craniofacial Precision Medicine Research, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China Department of General Dentistry, College of Stomatology, Xi’an Jiaotong University, Xi’an, Shaanxi, China

Collapse

Investigating causal associations among gut microbiota, metabolites, and liver diseases: a Mendelian randomization study

Objective

There is some evidence for an association between gut microbiota and nonalcoholic fatty liver disease (NAFLD), alcoholic liver disease (ALD), and viral hepatitis, but no studies have explored their causal relationship.

Methods

Instrumental variables of the gut microbiota (N = 13266) and gut microbiota-derived metabolites (N = 7824) were acquired, and a Mendelian randomization study was performed to explore their influence on NAFLD (1483 European cases and 17,781 European controls), ALD (2513 European cases and 332,951 European controls), and viral hepatitis risk (1971 European cases and 340,528 European controls). The main method for examining causality is inverse variance weighting (IVW).

Results

IVW results confirmed that Anaerotruncus (p = 0.0249), Intestinimonas (p = 0.0237), Lachnoclostridium (p = 0.0245), Lachnospiraceae NC2004 group (p = 0.0083), Olsenella (p = 0.0163), and Peptococcus (p = 0.0472) were protective factors for NAFLD, and Ruminococcus 1 (p = 0.0120) was detrimental for NAFLD. The higher abundance of three genera, Lachnospira (p = 0.0388), Desulfovibrio (p = 0.0252), and Ruminococcus torques group (p = 0.0364), was correlated with a lower risk of ALD, while Ruminococcaceae UCG 002 level was associated with a higher risk of ALD (p = 0.0371). The Alistipes (p = 0.0069) and Ruminococcaceae NK4A214 group (p = 0.0195) were related to a higher risk of viral hepatitis. Besides, alanine (p = 0.0076) and phenyllactate (p = 0.0100) were found to be negatively correlated with NAFLD, while stachydrine (Op = 0.0244) was found to be positively associated with NAFLD. The phenylacetate (p = 0.0353) and ursodeoxycholate (p = 0.0144) had a protective effect on ALD, while the threonate (p = 0.0370) exerted a detrimental influence on ALD. The IVW estimates of alanine (p = 0.0408) and cholate (p = 0.0293) showed their suggestive harmful effects against viral hepatitis, while threonate (p = 0.0401) displayed its suggestive protective effect against viral hepatitis.

Conclusion

In conclusion, our research supported causal links between the gut microbiome and its metabolites and NAFLD, ALD, and viral hepatitis.

Causal Relationship Between Basal Metabolic Rate and Alzheimer's Disease: A Bidirectional Two-sample Mendelian Randomization Study

INTRODUCTION

Objective observational studies have shown that basal metabolic rate (BMR) decreases in patients with Alzheimer's disease (AD), but the causal relationship between BMR and AD has not been established. We determined the causal relationship between BMR and AD by two-way Mendelian randomization (MR) and investigated the impact of factors associated with BMR on AD.

METHODS

We obtained BMR (n = 454,874) and AD from a large genome-wide association study (GWAS) database (21,982 patients with AD, 41,944 controls). The causal relationship between AD and BMR was investigated using two-way MR. Additionally, we identified the causal relationship between AD and factors related with BMR, hyperthyroidism (hy/thy) and type 2 diabetes (T2D), height and weight.

RESULTS

BMR had a causal relationship with AD [451 single nucleotide polymorphisms (SNPs), odds ratio (OR) 0.749, 95% confidence intervals (CIs) 0.663-0.858, P = 2.40E-03]. There was no causal relationship between hy/thy or T2D and AD (P > 0.05). The bidirectional MR showed that there was also a causal relationship between AD and BMR (OR 0.992, Cls 0.987-0.997, N_SNPs18, P = 1.50E-03). BMR, height and weight have a protective effect on AD. Based on MVMR analysis, we found that genetically determined height and weight may be adjusted by BMR to have a causal effect on AD, not height and weight themselves.

CONCLUSION

Our study showed that higher BMR reduced the risk of AD, and patients with AD had a lower BMR. Because of a positive correlation with BMR, height and weight may have a protective effect on AD. The two metabolism-related diseases, hy/thy and T2D, had no causal relationship with AD.

Body Shape Phenotypes and Breast Cancer Risk: A Mendelian Randomization Analysis

Observational and genetic studies have linked different anthropometric traits to breast cancer (BC) risk, with inconsistent results. We aimed to investigate the association between body shape defined by a principal component (PC) analysis of anthropometric traits (body mass index [BMI], height, weight, waist-to-hip ratio [WHR], and waist and hip circumference) and overall BC risk and by tumor sub-type (luminal A, luminal B, HER2+, triple negative, and luminal B/HER2 negative). We performed two-sample Mendelian randomization analyses to assess the association between 188 genetic variants robustly linked to the first three PCs and BC (133,384 cases/113,789 controls from the Breast Cancer Association Consortium (BCAC)). PC1 (general adiposity) was inversely associated with overall BC risk (0.89 per 1 SD [95% CI: 0.81-0.98]; p-value = 0.016). PC2 (tall women with low WHR) was weakly positively associated with overall BC risk (1.05 [95% CI: 0.98-1.12]; p-value = 0.135), but with a confidence interval including the null. PC3 (tall women with large WHR) was not associated with overall BC risk. Some of these associations differed by BC sub-types. For instance, PC2 was positively associated with a risk of luminal A BC sub-type (1.09 [95% CI: 1.01-1.18]; p-value = 0.02). To clarify the inverse association of PC1 with breast cancer risk, future studies should examine independent risk associations of this body shape during childhood/adolescence and adulthood.

Education plays a crucial role in the pathway from poverty to smoking: a Mendelian randomization study

BACKGROUND AND AIMS

Disproportionately high rates of smoking have been found in low-income communities, but the causal direction and role of education in this relationship remains less well understood. Here, we used bidirectional Mendelian randomization (MR) to measure the causal relationships between smoking, income and education.

DESIGN

Two-sample univariable and multivariable MR analyses were conducted to evaluate the total and direct effect of income and education on tobacco smoking. The effects of smoking on education and income were explored with reverse MR analysis.

SETTING

European ancestry.

PARTICIPANTS

The most recent large-scale genome-wide association study (GWAS) summary data on educational attainment, household income and smoking (n = 143 210-766 345).

MEASUREMENTS

Genetic variants for exposures including income, education and smoking.

FINDINGS

Both income and education had protective effects against smoking, especially for smoking initiation (education: β = -0.447, 95% CI = -0.508 to -0.387, P < 0.001; income: β = -0.290, 95% CI = -0.43 to -0.149, P < 0.001) and cessation (education: β = -0.364, 95% CI = -0.429 to -0.298, P < 0.001; income: β = -0.323, 95% CI = -0.448 to -0.197, P < 0.001). Here, higher scores in cessation indicated a lower likelihood of quitting according to the coding scheme. There was little evidence that income influenced smoking once education was controlled for, whereas education could significantly affect smoking behaviours independently of income (P = 3.40 × 10^-10 -0.0272). The reverse MR results suggested that smoking may result in a loss of years of schooling (β = -0.190, 95% CI = -0.297 to -0.083, P < 0.001) and reduced earnings (β = -0.204, 95% CI = -0.347 to -0.060, P = 0.006).

CONCLUSIONS

Education appears to play an important role in the relationship between income and smoking. There is a bidirectional association of smoking with socioeconomic status.

Nonalcoholic fatty liver disease and cardiovascular diseases: A Mendelian randomization study

BACKGROUND

Evidence suggests that nonalcoholic fatty liver disease (NAFLD) is associated with cardiovascular diseases (CVDs). However, the results are inconsistent, and the causality remains to be established.

OBJECTIVE

We aimed to investigate the potential causal relationship between NAFLD and CVDs, including arterial stiffness, coronary artery disease, heart failure, stroke, ischemic stroke and its subtypes using two-sample Mendelian randomization (MR).

METHODS

Genetic instruments were used as proxies for NAFLD. Publicly available summary-level data were obtained from the UK Biobank, the CARDIoGRAMplusC4D Consortium, the MEGASTROKE Consortium, and other consortia. Six complementary MR methods were performed, including inverse variance weighted method (IVW), MR-Egger, weighted median, weighted mode, MR-PRESSO, and MR-RAPS.

RESULTS

NAFLD was significantly associated with arterial stiffness (β = 0.04 [95%CI, 0.02-0.06], P = 5.53E-04). Moreover, the results remained consistent and robust in the sensitivity analysis. As for heart failure, the IVW method suggested that NAFLD was significantly associated with heart failure (OR = 1.08, 95%CI: 1.02-1.14, P = 0.005) in the absence of pleiotropy. However, there were no significant associations of NAFLD with coronary artery disease, stroke, ischemic stroke, or any ischemic stroke subtype.

CONCLUSION

The MR study supported the causal effect of NAFLD on arterial stiffness. However, the study did not provide enough evidence suggesting the causal associations of NAFLD with heart failure, coronary artery disease, and any stroke subtypes.

Impact of nonrandom selection mechanisms on the causal effect estimation for two-sample Mendelian randomization methods

Nonrandom selection in one-sample Mendelian Randomization (MR) results in biased estimates and inflated type I error rates only when the selection effects are sufficiently large. In two-sample MR, the different selection mechanisms in two samples may more seriously affect the causal effect estimation. Firstly, we propose sufficient conditions for causal effect invariance under different selection mechanisms using two-sample MR methods. In the simulation study, we consider 49 possible selection mechanisms in two-sample MR, which depend on genetic variants (G), exposures (X), outcomes (Y) and their combination. We further compare eight pleiotropy-robust methods under different selection mechanisms. Results of simulation reveal that nonrandom selection in sample II has a larger influence on biases and type I error rates than those in sample I. Furthermore, selections depending on X+Y, G+Y, or G+X+Y in sample II lead to larger biases than other selection mechanisms. Notably, when selection depends on Y, bias of causal estimation for non-zero causal effect is larger than that for null causal effect. Especially, the mode based estimate has the largest standard errors among the eight methods. In the absence of pleiotropy, selections depending on Y or G in sample II show nearly unbiased causal effect estimations when the casual effect is null. In the scenarios of balanced pleiotropy, all eight MR methods, especially MR-Egger, demonstrate large biases because the nonrandom selections result in the violation of the Instrument Strength Independent of Direct Effect (InSIDE) assumption. When directional pleiotropy exists, nonrandom selections have a severe impact on the eight MR methods. Application demonstrates that the nonrandom selection in sample II (coronary heart disease patients) can magnify the causal effect estimation of obesity on HbA1c levels. In conclusion, nonrandom selection in two-sample MR exacerbates the bias of causal effect estimation for pleiotropy-robust MR methods.

It is well known that nonrandom selection in one-sample Mendelian Randomization (MR) can result in biased estimates and inflated type I error rates. Actually, two-sample MR analyses are more prone to be affected by nonrandom selection than one-sample MR analyses, because two samples for genome-wide association studies (GWAS) may be selected each under different mechanisms from the source population. Summary-level genetic association statistics in two-sample MR may be derived from different study designs such as case-control, case-only and cohort studies, which further inevitably affect the causal effect estimation of exposure on outcome. In this study, we firstly propose a theorem for causal effect invariance under different selection mechanisms. In the simulation, we design 49 combinations of nonrandom selection mechanisms in sample I and sample II, which are widespread in practical applications. The simulation results reveal that the selection mechanisms in sample II have a larger influence on biases and type I error rates than those in sample I. As an illustrative example, we find the nonrandom selection in sample II (coronary heart disease patients) can magnify the causal effect estimation of obesity on the HbA1c levels.

Genetics of substance use disorders in the era of big data

Substance use disorders (SUDs) are conditions in which the use of legal or illegal substances, such as nicotine, alcohol or opioids, results in clinical and functional impairment. SUDs and, more generally, substance use are genetically complex traits that are enormously costly on an individual and societal basis. The past few years have seen remarkable progress in our understanding of the genetics, and therefore the biology, of substance use and abuse. Various studies - including of well-defined phenotypes in deeply phenotyped samples, as well as broadly defined phenotypes in meta-analysis and biobank samples - have revealed multiple risk loci for these common traits. A key emerging insight from this work establishes a biological and genetic distinction between quantity and/or frequency measures of substance use (which may involve low levels of use without dependence), versus symptoms related to physical dependence.

No Casual Relationship Between T2DM and the Risk of Infectious Diseases: A Two-Sample Mendelian Randomization Study

Background

In epidemiological studies, it has been proven that the occurrence of type 2 diabetes mellitus (T2DM) is related to an increased risk of infectious diseases. However, it is still unclear whether the relationship is casual.

Methods

We employed a two-sample Mendelian randomization (MR) to clarify the causal effect of T2DM on high-frequency infectious diseases: sepsis, skin and soft tissue infections (SSTIs), urinary tract infections (UTIs), pneumonia, and genito-urinary infection (GUI) in pregnancy. And then, we analyzed the genome-wide association study (GWAS) meta-analysis of European-descent individuals and conducted T2DM-related single-nucleotide polymorphisms (SNPs) as instrumental variables (IVs) that were associated with genome-wide significance (p < 5 × 10^–8). MR estimates were obtained using the inverse variance-weighted (IVW), the MR-Egger regression, the simple mode (SM), weighted median, and weighted mode.

Results

The UK Biobank (UKB) cohort (n > 500,000) provided data for GWASs on infectious diseases. MR analysis showed little evidence of a causal relationship of T2DM with five mentioned infections’ (sepsis, SSTI, UTI, pneumonia, and GUI in pregnancy) susceptibility [odds ratio (OR) = 0.99999, p = 0.916; OR = 0.99986, p = 0.233; OR = 0.99973, p = 0.224; OR = 0.99997, p = 0.686; OR, 1.00002, p = 0.766]. Sensitivity analysis showed similar results, indicating the robustness of causality. There were no heterogeneity and pleiotropic bias.

Conclusion

T2DM would not be causally associated with high-frequency infectious diseases (including sepsis, SSTI, UTI, pneumonia, and GUI in pregnancy).

Assessing the genetic relationship between gastro-esophageal reflux disease and risk of COVID-19 infection

Background

Symptoms related with Gastro-esophageal reflux disease (GERD) were previously shown to be linked with increased risk for the 2019 coronavirus disease (COVID-19). We aim to interrogate the possibility of a shared genetic basis between GERD and COVID-19 outcomes.

Methods

Using published GWAS data for GERD (78 707 cases; 288 734 controls) and COVID-19 susceptibility (up to 32 494 cases; 1.5 million controls), we examined the genetic relationship between GERD and three COVID-19 outcomes: risk of developing severe COVID-19, COVID-19 hospitalization and overall COVID-19 risk. We estimated the genetic correlation between GERD and COVID-19 outcomes followed by Mendelian randomization (MR) analyses to assess genetic causality. Conditional analyses were conducted to examine whether known COVID-19 risk factors (obesity, smoking, type-II diabetes, coronary artery disease) can explain the relationship between GERD and COVID-19.

Results

We found small to moderate genetic correlations between GERD and COVID-19 outcomes (rg between 0.06–0.24). MR analyses revealed a OR of 1.15 (95% CI: 0.96–1.39) for severe COVID-19; 1.16 (1.01–1.34) for risk of COVID-19 hospitalization; 1.05 (0.97–1.13) for overall risk of COVID-19 per doubling of odds in developing GERD. The genetic correlation/associations between GERD and COVID-19 showed mild attenuation towards the null when obesity and smoking was adjusted for.

Conclusions

Susceptibility for GERD and risk of COVID-19 hospitalization were genetically correlated, with MR findings supporting a potential causal role between the two. The genetic association between GERD and COVID-19 was partially attenuated when obesity is accounted for, consistent with obesity being a major risk factor for both diseases.

Vitamin D and nonalcoholic fatty liver disease

PURPOSE OF REVIEW

Vitamin D deficiency may impact disease progression of nonalcoholic fatty liver disease (NAFLD). The aim of this work was to review recent studies examining either vitamin D status or the effects of supplementation in patients with NAFLD, along with investigating the roles of genetic polymorphisms and the gut microbiome.

RECENT FINDINGS

Six heterogeneous observational studies of vitamin D status, and four randomized controlled intervention trials of vitamin D supplementation in NAFLD were conflicting. All studies were hampered by the challenges of diagnosing NAFLD, were underpowered, and lacked data on clinically important outcomes. The results of three cross-sectional studies, including a Mendelian randomization study, provide limited evidence for a role for genetic modifiers of vitamin D status in NAFLD. Genetic and experimental evidence suggests that vitamin D and the vitamin D receptor (VDR) may influence the gut microbiome in health and disease.

SUMMARY

The evidence relating either lower vitamin D status to the prevalence and severity of NAFLD, or examining vitamin D supplementation in patients with NAFLD is inconclusive. Larger, higher quality trials with relevant endpoints are needed. Further mechanistic studies on the roles of vitamin D and VDR in influencing the gut-liver axis in NAFLD are warranted.

New alcohol-related genes suggest shared genetic mechanisms with neuropsychiatric disorders

Excessive alcohol consumption is one of the main causes of death and disability worldwide. Alcohol consumption is a heritable complex trait. Here we conducted a meta-analysis of genome-wide association studies of alcohol consumption (g d-1) from the UK Biobank, the Alcohol Genome-Wide Consortium and the Cohorts for Heart and Aging Research in Genomic Epidemiology Plus consortia, collecting data from 480,842 people of European descent to decipher the genetic architecture of alcohol intake. We identified 46 new common loci and investigated their potential functional importance using magnetic resonance imaging data and gene expression studies. We identify genetic pathways associated with alcohol consumption and suggest genetic mechanisms that are shared with neuropsychiatric disorders such as schizophrenia.

Association between blood pressure and risk of cancer development: a systematic review and meta-analysis of observational studies

With the exception of renal cell carcinoma, studies assessing the association between hypertension and other cancers are inconsistent. We conducted a meta-analysis to assess this evidence. We included observational studies investigating the association between any definition of hypertension or systolic and diastolic blood pressure and risk of any cancer, after searching PubMed until November 2017. We calculated summary relative risks (RR) and 95% confidence intervals (CI) using inverse-variance weighted random effects methods. A total of 148 eligible publications were identified out of 39,891 initially screened citations. Considering only evidence from 85 prospective studies, positive associations were observed between hypertension and kidney, colorectal and breast cancer. Positive associations between hypertension and risk of oesophageal adenocarcinoma and squamous cell carcinoma, liver and endometrial cancer were also observed, but the majority of studies did not perform comprehensive multivariable adjustments. Systolic and diastolic blood pressure were positively associated with risk of kidney cancer but not with other cancers. In addition to the previously well-described association between hypertension and risk of kidney cancer, the current meta-analysis suggested that hypertensive individuals may also be at higher risk of colorectal and breast cancer. However, careful interpretation is required as most meta-analyses included relatively small number of studies, several relative risks had weak or moderate magnitude and maybe affected by residual confounding.