Figure 1: The genome-proteome-disease network.

(a) Data sources integrated into the network, indicating the number and type of the overlapping associations, from the SNP to the disease end point; all associations are freely accessible at http://proteomics.gwas.eu. (b) Circular plot of all cis- and trans-associations, cis-pQTLs are indicated by triangles, trans-pQTLs connect associated variant locations and trans-encoded protein locations. An interactive version of this circular plot constitutes an entry point to query the integrated web-server. (c) Example of a genome-proteome-disease sub-network obtained from the server for a query using the search word ‘Crohn’s Disease’. Network elements are disease traits (pink hexagons), pQTL loci (green diamonds), protein levels (blue ovals); nodes are connected by genetic associations, partial correlations and disease GWAS associations. This example (edited here for clarity) revealed four risk loci that associated with plasma levels of C7, MST, IL23R and IL18R, respectively. These four proteins all have a major role in auto-immune disorders. Partial correlations between neighbouring proteins reveal pathways that may be involved in the aetiology of Crohn’s disease. Similar networks can be retrieved starting with a query using any of the 539 pQTLs, 1,124 proteins and 42 unique co-associated disease end points. In the integrated web-server, all items are interactively linked to association data from the discovery and the replication study, regional association plots based on imputed variants, locus annotations including co-associated eQTL-, meQTL-, mQTL-, regulatory-, coding- and disease risk-variants, and link-outs to relevant protein databases, original data sources and primary publications. The links in this network reflect the outcome of many natural experiments, represented by genetic variations observed in the genomes of hundreds of individuals from the general population and probed by deep proteomics phenotyping using over 1,100 different aptamers.