Linking protein to phenotype with Mendelian Randomization detects 38 proteins with causal roles in human diseases and traits.

Andrew D Bretherick,Peter K Joshi,Chris Haley,Chris P Ponting,James F Wilson,Pau Navarro,Alan F Wright,Harry Campbell,David W Clark,Igor Rudan,Caroline Hayward,Carmen Amador,Yanni Zeng,Oriol Canela-Xandri,Konrad Rawlik,Albert Tenesa,Thibaud Boutin ,Véronique Vitart ,J Kenneth Baillie

doi:10.1371/journal.pgen.1008785

Abstract

To efficiently transform genetic associations into drug targets requires evidence that a particular gene, and its encoded protein, contribute causally to a disease. To achieve this, we employ a three-step proteome-by-phenome Mendelian Randomization (MR) approach. In step one, 154 protein quantitative trait loci (pQTLs) were identified and independently replicated. From these pQTLs, 64 replicated locally-acting variants were used as instrumental variables for proteome-by-phenome MR across 846 traits (step two). When its assumptions are met, proteome-by-phenome MR, is equivalent to simultaneously running many randomized controlled trials. Step 2 yielded 38 proteins that significantly predicted variation in traits and diseases in 509 instances. Step 3 revealed that amongst the 271 instances from GeneAtlas (UK Biobank), 77 showed little evidence of pleiotropy (HEIDI), and 92 evidence of colocalization (eCAVIAR). Results were wide ranging: including, for example, new evidence for a causal role of tyrosine-protein phosphatase non-receptor type substrate 1 (SHPS1; SIRPA) in schizophrenia, and a new finding that intestinal fatty acid binding protein (FABP2) abundance contributes to the pathogenesis of cardiovascular disease. We also demonstrated confirmatory evidence for the causal role of four further proteins (FGF5, IL6R, LPL, LTA) in cardiovascular disease risk.

Highlights

An initial goal of drug development is the identification of targets—in most cases, proteins— whose interaction with a drug ameliorates the development, progression, or symptoms of disease
Linking protein to phenotype with Mendelian Randomization
Of the 209 lead-SNPs identified in the discovery cohort at this threshold, 154 were successfully replicated. These represented pQTLs for 82 proteins, all but two proteins were successfully mapped to an autosomal gene (Ensembl GRCh37; ‘start_position: start position (GRCh37))

Summary

Introduction

An initial goal of drug development is the identification of targets—in most cases, proteins— whose interaction with a drug ameliorates the development, progression, or symptoms of disease. A large proportion of drugs fail during the last stages of development—clinical trials—because their targets do not alter whole-organism phenotypes as expected from observational and other pre-clinical research [2]. Genetic approaches to drug development [3] offer a distinct advantage over observational studies. It is estimated that by selecting targets with genetic evidence, the chance of success of those targets doubles in subsequent clinical development [4]. A recent study found that 12% of all targets for licensed drugs could be rediscovered using GWA studies [5]. There have been a number of recent high-profile successes prioritizing therapeutic targets at genome-wide scales [6,7]. The genetic associations of disease are often still not immediately interpretable [8] and many disease-associated variants alter protein levels via poorly understood mechanisms

Methods

Results

Conclusion