Philip Haycock

Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator

Developments in genome‐wide association studies and the increasing availability of summary genetic association data have made application of Mendelian randomization relatively straightforward. However, obtaining reliable results from a Mendelian randomization investigation remains problematic, as the conventional inverse‐variance weighted method only gives consistent estimates if all of the genetic variants in the analysis are valid instrumental variables. We present a novel weighted median estimator for combining data on multiple genetic variants into a single causal estimate. This estimator is consistent even when up to 50% of the information comes from invalid instrumental variables. In a simulation analysis, it is shown to have better finite‐sample Type 1 error rates than the inverse‐variance weighted method, and is complementary to the recently proposed MR‐Egger (Mendelian randomization‐Egger) regression method. In analyses of the causal effects of low‐density lipoprotein cholesterol and high‐density lipoprotein cholesterol on coronary artery disease risk, the inverse‐variance weighted method suggests a causal effect of both lipid fractions, whereas the weighted median and MR‐Egger regression methods suggest a null effect of high‐density lipoprotein cholesterol that corresponds with the experimental evidence. Both median‐based and MR‐Egger regression methods should be considered as sensitivity analyses for Mendelian randomization investigations with multiple genetic variants.

LD Hub: a centralized database and web interface to perform LD score regression that maximizes the potential of summary level GWAS data for SNP heritability and genetic correlation analysis

Motivation: LD score regression is a reliable and efficient method of using genome-wide association study (GWAS) summary-level results data to estimate the SNP heritability of complex traits and diseases, partition this heritability into functional categories, and estimate the genetic correlation between different phenotypes. Because the method relies on summary level results data, LD score regression is computationally tractable even for very large sample sizes. However, publicly available GWAS summary-level data are typically stored in different databases and have different formats, making it difficult to apply LD score regression to estimate genetic correlations across many different traits simultaneously. Results: In this manuscript, we describe LD Hub - a centralized database of summary-level GWAS results for 173 diseases/traits from different publicly available resources/consortia and a web interface that automates the LD score regression analysis pipeline. To demonstrate functionality and validate our software, we replicated previously reported LD score regression analyses of 49 traits/diseases using LD Hub; and estimated SNP heritability and the genetic correlation across the different phenotypes. We also present new results obtained by uploading a recent atopic dermatitis GWAS meta-analysis to examine the genetic correlation between the condition and other potentially related traits. In response to the growing availability of publicly accessible GWAS summary-level results data, our database and the accompanying web interface will ensure maximal uptake of the LD score regression methodology, provide a useful database for the public dissemination of GWAS results, and provide a method for easily screening hundreds of traits for overlapping genetic aetiologies. Availability and Implementation: The web interface and instructions for using LD Hub are available at http://ldsc.broadinstitute.org/ Contact: jie.zheng@bristol.ac.uk Supplementary information: Supplementary data are available at Bioinformatics online.

Obesity, metabolic factors and risk of different histological types of lung cancer: A Mendelian randomization study.

Background Assessing the relationship between lung cancer and metabolic conditions is challenging because of the confounding effect of tobacco. Mendelian randomization (MR), or the use of genetic instrumental variables to assess causality, may help to identify the metabolic drivers of lung cancer. Methods and findings We identified genetic instruments for potential metabolic risk factors and evaluated these in relation to risk using 29,266 lung cancer cases (including 11,273 adenocarcinomas, 7,426 squamous cell and 2,664 small cell cases) and 56,450 controls. The MR risk analysis suggested a causal effect of body mass index (BMI) on lung cancer risk for two of the three major histological subtypes, with evidence of a risk increase for squamous cell carcinoma (odds ratio (OR) [95% confidence interval (CI)] = 1.20 [1.01–1.43] and for small cell lung cancer (OR [95%CI] = 1.52 [1.15–2.00]) for each standard deviation (SD) increase in BMI [4.6 kg/m2]), but not for adenocarcinoma (OR [95%CI] = 0.93 [0.79–1.08]) (Pheterogeneity = 4.3x10-3). Additional analysis using a genetic instrument for BMI showed that each SD increase in BMI increased cigarette consumption by 1.27 cigarettes per day (P = 2.1x10-3), providing novel evidence that a genetic susceptibility to obesity influences smoking patterns. There was also evidence that low-density lipoprotein cholesterol was inversely associated with lung cancer overall risk (OR [95%CI] = 0.90 [0.84–0.97] per SD of 38 mg/dl), while fasting insulin was positively associated (OR [95%CI] = 1.63 [1.25–2.13] per SD of 44.4 pmol/l). Sensitivity analyses including a weighted-median approach and MR-Egger test did not detect other pleiotropic effects biasing the main results. Conclusions Our results are consistent with a causal role of fasting insulin and low-density lipoprotein cholesterol in lung cancer etiology, as well as for BMI in squamous cell and small cell carcinoma. The latter relation may be mediated by a previously unrecognized effect of obesity on smoking behavior.

Hypomethylation of smoking-related genes is associated with future lung cancer in four prospective cohorts

DNA hypomethylation in certain genes is associated with tobacco exposure but it is unknown whether these methylation changes translate into increased lung cancer risk. In an epigenome-wide study of DNA from pre-diagnostic blood samples from 132 case–control pairs in the NOWAC cohort, we observe that the most significant associations with lung cancer risk are for cg05575921 in AHRR (OR for 1 s.d.=0.37, 95% CI: 0.31–0.54, P-value=3.3 × 10−11) and cg03636183 in F2RL3 (OR for 1 s.d.=0.40, 95% CI: 0.31–0.56, P-value=3.9 × 10−10), previously shown to be strongly hypomethylated in smokers. These associations remain significant after adjustment for smoking and are confirmed in additional 664 case–control pairs tightly matched for smoking from the MCCS, NSHDS and EPIC HD cohorts. The replication and mediation analyses suggest that residual confounding is unlikely to explain the observed associations and that hypomethylation of these CpG sites may mediate the effect of tobacco on lung cancer risk.

DNA methylation changes measured in pre-diagnostic peripheral blood samples are associated with smoking and lung cancer risk.

DNA methylation changes are associated with cigarette smoking. We used the Illumina Infinium HumanMethylation450 array to determine whether methylation in DNA from pre‐diagnostic, peripheral blood samples is associated with lung cancer risk. We used a case‐control study nested within the EPIC‐Italy cohort and a study within the MCCS cohort as discovery sets (a total of 552 case‐control pairs). We validated the top signals in 429 case‐control pairs from another 3 studies. We identified six CpGs for which hypomethylation was associated with lung cancer risk: cg05575921 in the AHRR gene (p‐valuepooled = 4 × 10−17), cg03636183 in the F2RL3 gene (p‐valuepooled = 2 × 10 − 13), cg21566642 and cg05951221 in 2q37.1 (p‐valuepooled = 7 × 10−16 and 1 × 10−11 respectively), cg06126421 in 6p21.33 (p‐valuepooled = 2 × 10−15) and cg23387569 in 12q14.1 (p‐valuepooled = 5 × 10−7). For cg05951221 and cg23387569 the strength of association was virtually identical in never and current smokers. For all these CpGs except for cg23387569, the methylation levels were different across smoking categories in controls (p‐valuesheterogeneity ≤ 1.8 x10 − 7), were lowest for current smokers and increased with time since quitting for former smokers. We observed a gain in discrimination between cases and controls measured by the area under the ROC curve of at least 8% (p‐values ≥ 0.003) in former smokers by adding methylation at the 6 CpGs into risk prediction models including smoking status and number of pack‐years. Our findings provide convincing evidence that smoking and possibly other factors lead to DNA methylation changes measurable in peripheral blood that may improve prediction of lung cancer risk.

Circulating vitamin D concentration and risk of seven cancers: Mendelian randomisation study

Objective To determine if circulating concentrations of vitamin D are causally associated with risk of cancer. Design Mendelian randomisation study. Setting Large genetic epidemiology networks (the Genetic Associations and Mechanisms in Oncology (GAME-ON), the Genetic and Epidemiology of Colorectal Cancer Consortium (GECCO), and the Prostate Cancer Association Group to Investigate Cancer Associated Alterations in the Genome (PRACTICAL) consortiums, and the MR-Base platform). Participants 70 563 cases of cancer (22 898 prostate cancer, 15 748 breast cancer, 12 537 lung cancer, 11 488 colorectal cancer, 4369 ovarian cancer, 1896 pancreatic cancer, and 1627 neuroblastoma) and 84 418 controls. Exposures Four single nucleotide polymorphisms (rs2282679, rs10741657, rs12785878 and rs6013897) associated with vitamin D were used to define a multi-polymorphism score for circulating 25-hydroxyvitamin D (25(OH)D) concentrations. Main outcomes measures The primary outcomes were the risk of incident colorectal, breast, prostate, ovarian, lung, and pancreatic cancer and neuroblastoma, which was evaluated with an inverse variance weighted average of the associations with specific polymorphisms and a likelihood based approach. Secondary outcomes based on cancer subtypes by sex, anatomic location, stage, and histology were also examined. Results There was little evidence that the multi-polymorphism score of 25(OH)D was associated with risk of any of the seven cancers or their subtypes. Specifically, the odds ratios per 25 nmol/L increase in genetically determined 25(OH)D concentrations were 0.92 (95% confidence interval 0.76 to 1.10) for colorectal cancer, 1.05 (0.89 to 1.24) for breast cancer, 0.89 (0.77 to 1.02) for prostate cancer, and 1.03 (0.87 to 1.23) for lung cancer. The results were consistent with the two different analytical approaches, and the study was powered to detect relative effect sizes of moderate magnitude (for example, 1.20-1.50 per 25 nmol/L decrease in 25(OH)D for most primary cancer outcomes. The Mendelian randomisation assumptions did not seem to be violated. Conclusions There is little evidence for a linear causal association between circulating vitamin D concentration and risk of various types of cancer, though the existence of causal clinically relevant effects of low magnitude cannot be ruled out. These results, in combination with previous literature, provide evidence that population-wide screening for vitamin D deficiency and subsequent widespread vitamin D supplementation should not currently be recommended as a strategy for primary cancer prevention.

Apolipoprotein(a) isoform size, lipoprotein(a) concentration, and coronary artery disease: a mendelian randomisation analysis

Summary Background The lipoprotein(a) pathway is a causal factor in coronary heart disease. We used a genetic approach to distinguish the relevance of two distinct components of this pathway, apolipoprotein(a) isoform size and circulating lipoprotein(a) concentration, to coronary heart disease. Methods In this mendelian randomisation study, we measured lipoprotein(a) concentration and determined apolipoprotein(a) isoform size with a genetic method (kringle IV type 2 [KIV2] repeats in the LPA gene) and a serum-based electrophoretic assay in patients and controls (frequency matched for age and sex) from the Pakistan Risk of Myocardial Infarction Study (PROMIS). We calculated odds ratios (ORs) for myocardial infarction per 1-SD difference in either LPA KIV2 repeats or lipoprotein(a) concentration. In a genome-wide analysis of up to 17 503 participants in PROMIS, we identified genetic variants associated with either apolipoprotein(a) isoform size or lipoprotein(a) concentration. Using a mendelian randomisation study design and genetic data on 60 801 patients with coronary heart disease and 123 504 controls from the CARDIoGRAMplusC4D consortium, we calculated ORs for myocardial infarction with variants that produced similar differences in either apolipoprotein(a) isoform size in serum or lipoprotein(a) concentration. Finally, we compared phenotypic versus genotypic ORs to estimate whether apolipoprotein(a) isoform size, lipoprotein(a) concentration, or both were causally associated with coronary heart disease. Findings The PROMIS cohort included 9015 patients with acute myocardial infarction and 8629 matched controls. In participants for whom KIV2 repeat and lipoprotein(a) data were available, the OR for myocardial infarction was 0·93 (95% CI 0·90–0·97; p<0·0001) per 1-SD increment in LPA KIV2 repeats after adjustment for lipoprotein(a) concentration and conventional lipid concentrations. The OR for myocardial infarction was 1·10 (1·05–1·14; p<0·0001) per 1-SD increment in lipoprotein(a) concentration, after adjustment for LPA KIV2 repeats and conventional lipids. Genome-wide analysis identified rs2457564 as a variant associated with smaller apolipoprotein(a) isoform size, but not lipoprotein(a) concentration, and rs3777392 as a variant associated with lipoprotein(a) concentration, but not apolipoprotein(a) isoform size. In 60 801 patients with coronary heart disease and 123 504 controls, OR for myocardial infarction was 0·96 (0·94–0·98; p<0·0001) per 1-SD increment in apolipoprotein(a) protein isoform size in serum due to rs2457564, which was directionally concordant with the OR observed in PROMIS for a similar change. The OR for myocardial infarction was 1·27 (1·07–1·50; p=0·007) per 1-SD increment in lipoprotein(a) concentration due to rs3777392, which was directionally concordant with the OR observed for a similar change in PROMIS. Interpretation Human genetic data suggest that both smaller apolipoprotein(a) isoform size and increased lipoprotein(a) concentration are independent and causal risk factors for coronary heart disease. Lipoprotein(a)-lowering interventions could be preferentially effective in reducing the risk of coronary heart disease in individuals with smaller apolipoprotein(a) isoforms. Funding British Heart Foundation, US National Institutes of Health, Fogarty International Center, Wellcome Trust, UK Medical Research Council, UK National Institute for Health Research, and Pfizer.

The Role of Obesity, Type 2 Diabetes, and Metabolic Factors in Pancreatic Cancer: A Mendelian Randomization Study

Abstract Background Risk factors for pancreatic cancer include a cluster of metabolic conditions such as obesity, hypertension, dyslipidemia, insulin resistance, and type 2 diabetes. Given that these risk factors are correlated, separating out causal from confounded effects is challenging. Mendelian randomization (MR), or the use of genetic instrumental variables, may facilitate the identification of the metabolic drivers of pancreatic cancer. Methods We identified genetic instruments for obesity, body shape, dyslipidemia, insulin resistance, and type 2 diabetes in order to evaluate their causal role in pancreatic cancer etiology. These instruments were analyzed in relation to risk using a likelihood-based MR approach within a series of 7110 pancreatic cancer patients and 7264 control subjects using genome-wide data from the Pancreatic Cancer Cohort Consortium (PanScan) and the Pancreatic Cancer Case-Control Consortium (PanC4). Potential unknown pleiotropic effects were assessed using a weighted median approach and MR-Egger sensitivity analyses. Results Results indicated a robust causal association of increasing body mass index (BMI) with pancreatic cancer risk (odds ratio [OR] = 1.34, 95% confidence interval [CI] = 1.09 to 1.65, for each standard deviation increase in BMI [4.6 kg/m2]). There was also evidence that genetically increased fasting insulin levels were causally associated with an increased risk of pancreatic cancer (OR = 1.66, 95% CI = 1.05 to 2.63, per SD [44.4 pmol/L]). Notably, no evidence of a causal relationship was observed for type 2 diabetes, nor for dyslipidemia. Sensitivity analyses did not indicate that pleiotropy was an important source of bias. Conclusions Our results suggest a causal role of BMI and fasting insulin in pancreatic cancer etiology.

Recent Developments in Mendelian Randomization Studies

Purpose of ReviewMendelian randomization (MR) is a strategy for evaluating causality in observational epidemiological studies. MR exploits the fact that genotypes are not generally susceptible to reverse causation and confounding, due to their fixed nature and Mendel’s First and Second Laws of Inheritance. MR has the potential to provide information on causality in many situations where randomized controlled trials are not possible, but the results of MR studies must be interpreted carefully to avoid drawing erroneous conclusions.Recent FindingsIn this review, we outline the principles behind MR, as well as assumptions and limitations of the method. Extensions to the basic approach are discussed, including two-sample MR, bidirectional MR, two-step MR, multivariable MR, and factorial MR. We also consider some new applications and recent developments in the methodology, including its ability to inform drug development, automation of the method using tools such as MR-Base, and phenome-wide and hypothesis-free MR.SummaryIn conjunction with the growing availability of large-scale genomic databases, higher level of automation and increased robustness of the methods, MR promises to be a valuable strategy to examine causality in complex biological/omics networks, inform drug development and prioritize intervention targets for disease prevention in the future.

The causal relevance of body mass index in different histological types of lung cancer: A Mendelian randomization study

Body mass index (BMI) is inversely associated with lung cancer risk in observational studies, even though it increases the risk of several other cancers, which could indicate confounding by tobacco smoking or reverse causality. We used the two-sample Mendelian randomization (MR) approach to circumvent these limitations of observational epidemiology by constructing a genetic instrument for BMI, based on results from the GIANT consortium, which was evaluated in relation to lung cancer risk using GWAS results on 16,572 lung cancer cases and 21,480 controls. Results were stratified by histological subtype, smoking status and sex. An increase of one standard deviation (SD) in BMI (4.65 Kg/m2) raised the risk for lung cancer overall (OR = 1.13; P = 0.10). This was driven by associations with squamous cell (SQ) carcinoma (OR = 1.45; P = 1.2 × 10−3) and small cell (SC) carcinoma (OR = 1.81; P = 0.01). An inverse trend was seen for adenocarcinoma (AD) (OR = 0.82; P = 0.06). In stratified analyses, a 1 SD increase in BMI was inversely associated with overall lung cancer in never smokers (OR = 0.50; P = 0.02). These results indicate that higher BMI may increase the risk of certain types of lung cancer, in particular SQ and SC carcinoma.