Identification of inflammatory bowel disease-related proteins using a reverse k-nearest neighbor search

Inflammatory bowel disease (IBD) is a chronic disease whose incidence and prevalence increase every year; however, the pathogenesis of IBD is still unclear. Thus, identifying IBD-related proteins is important for understanding its complex disease mechanism. Here, we propose a new and simple network-based approach using a reverse k-nearest neighbor (RkNN) search to identify novel IBD-related proteins. Protein-protein interactions (PPI) and Genome-Wide Association Studies (GWAS) were used in this study. After constructing the PPI network, the RkNN search was applied to all of the proteins to identify sets of influenced proteins among their k-nearest neighbors (kNNs). An observed protein whose influenced proteins were mostly known IBD-related proteins was statistically identified as a novel IBD-related protein. Our method outperformed a random aspect, kNN search, and centrality measures based on the network topology. A total of 39 proteins were identified as IBD-related proteins. Of these proteins, 71% were reported at least once in the literature as related to IBD. Additionally, these proteins were found over-represented in the IBD pathway and enriched in importantly functional pathways in IBD. In conclusion, the RkNN search with the statistical enrichment test is a great tool to identify IBD-related proteins to better understand its complex disease mechanism.