Abstract
With advances in genotyping, rich phenotyping, the use of electronic health records and the availability of large population-based cohorts, genome-wide association studies (GWAS) enable the identification of associations for complex traits at an unprecedented rate. Consequently, the landscape of post-GWAS analyses has also rapidly evolved to accommodate the increasing complexity and diversity of genetic research needs. Existing resources can be broadly classified into tools for functional annotation and interpretation, data dissemination platforms and visualization frameworks such as PheWeb21 (for related work, see Supplementary Information). We identified several important roadblocks in the current tool landscape. First, there is a lack of tools to easily process GWAS results into an interactive, user-friendly and performant web platform. Second, most existing tools limit analyses to single traits and focus on displaying information on specific variants. However, single-variant analyses suffer from reduced statistical power due to stringent multiple-testing corrections. By contrast, gene-based approaches aggregate variant-level signals within genes or genomic regions into composite gene-level statistics, thereby improving power to detect disease contributors. Third, none of the existing tools leverages the fact that disease-associated genes tend to cluster within specific regions of biological networks, forming disease modules that capture underlying pathogenic mechanisms and enable mining for hypothetical drug repurposing candidates targeting both the disease-associated genes and their functionally connected network neighbors.