Efficient candidate drug target discovery through proteogenomics in a Scottish cohort.

Understanding the genomic basis of human proteomic variability provides powerful tools to probe potential causal relationships of proteins and disease risk, and thus to prioritise candidate drug targets. Here, we investigated 6432 plasma proteins (1533 previously unstudied in large-scale proteomic GWAS) using the SomaLogic (v4.1) aptamer-based technology in a Scottish population from the Viking Genes study. A total of 505 significant independent protein quantitative trait loci (pQTL) were found for 455 proteins in blood plasma: 382 cis- (P < 5×10) and 123 trans- (P < 6.6×10). Of these, 31 cis-pQTL were for proteins with no previous GWAS. We leveraged these pQTL to perform causal inference using bidirectional Mendelian randomisation and colocalisation against complex traits of biomedical importance. We discovered 43 colocalising associations (with a posterior probability > 80% that pQTL and complex traits share a causal variant), pointing to plausible causal roles for the proteins. These findings include hitherto undiscovered causal links of leukocyte receptor tyrosine kinase (LTK) to type-2 diabetes and beta-1,3-glucuronyltransferase (B3GAT1) to prostate cancer. These new connections will help guide the search for new or repurposed therapies. Our findings provide strong support for continuing to increase the number of proteins studied using GWAS.