Mapping genetic determinants of 184 circulating proteins in 26,494 individuals to connect proteins and diseases

Abstract

We performed the largest genome-wide meta-analysis (GWAMA) (Max N=26,494) of the levels of 184 cardiovascular-related plasma protein levels to date and reported 592 independent loci (pQTL) associated with the level of at least one protein (1308 significant associations, median 6 per protein). We estimated that only between 8-37% of testable pQTL overlap with established expression quantitative trait loci (eQTL) using multiple methods, while 132 out of 1064 lead variants show evidence for transcription factor binding, and found that 75% of our pQTL are known DNA methylation QTL. We highlight the variation in genetic architecture between proteins and that proteins share genetic architecture with cardiometabolic complex traits. Using cis-instrument Mendelian randomisation (MR), we infer causal relationships for 11 proteins, recapitulating the previously reported relationship between PCSK9 and LDL cholesterol, replicating previous pQTL MR findings and discovering 16 causal relationships between protein levels and disease. Our MR results highlight IL2-RA as a candidate for drug repurposing for Crohn’s Disease as well as 2 novel therapeutic targets: IL-27 (Crohn’s disease) and TNFRSF14 (Inflammatory bowel disease, Multiple sclerosis and Ulcerative colitis). We have demonstrated the discoveries possible using our pQTL and highlight the potential of this work as a resource for genetic epidemiology.

Efficient candidate drug target discovery through proteogenomics in a Scottish cohort

“…As in other studies 9, 18 , an inverse relationship between the minor allele frequency and the absolute effect size was observed for both cis and trans associations. Overall, trans associations displayed both smaller effect sizes and were less detectable at lower allele frequencies (Fig.…”

“…Clearly, an increased sample size would enable discovery of variants with smaller effects. For instance, in this study only 4.6% of the proteins had a genome-wide significant cis association, while the largest proteomic studies report >90% 18 . Increasing the sample size would help not only with identifying smaller cis associations, but would also allow detection and pathway analysis of more trans associations, as their number increases and exceeds that of cis associations as the study power grows 15 .…”

Genetic regulation of the human plasma proteome in 54,306 UK Biobank participants

“…We estimated SNP-based heritability as a sum of contributions from significant lead pQTLs (pQTL component) and the remaining SNPs across the genome (excluding the pQTL region), which assumes a polygenic model (polygenic component) using the approach described in 36 ( Supplementary Table 11, Methods) . The mean total SNP-based heritability was 0.18 (5-95 th quantiles: 0.02-0.44) ( Figure 1d ).…”

“…For stroke, we found significant causal associations with increased risk of large artery ischaemic stroke subtype (MRlog(IS-LA OR)=0.27, p=0.011). Whilst genetic PCSK9 effects on LDL, total cholesterol and CHD have been found previously 36,75 , effects of PCSK9 on HDL cholesterol and large artery ischaemic stroke have not been substantiated by previous MR studies, likely due to decreased power. These findings extends the corroborated effects observed across multiple randomised clinical trials of PCSK9 inhibitors 72 .…”

Mass spectrometry‐based high‐throughput proteomics and its role in biomedical studies and systems biology

“…Whereas genomic quantitative trait loci (QTL) simply link genetic variants with phenotypes and diseases, pQTLs enable insights into the molecular mechanisms underpinning these associations, for example, by pinpointing the affected proteins and disentangling effects on protein dosage from protein structure [158]. Partly due to their accessibility, the analysis of human biofluids such as plasma have proven promising, for example, yielding a comprehensive plasma proteo-genomic map of human health and disease [156,159,160]. Knowledge that a genetic variant links to a disease, and is associated with changing protein levels is particularly valuable to establish target-indication pairs, and thus for the mining of drug targets [161].…”