Uncovering RNA-Binding Proteins Implicated in Human Cancers by Integrating Genomics with Network-Based Approaches

Abstract

RNA binding proteins (RBPs) are key regulatory mechanism in the cell. Their functions are varied depending on their localization, but many have been identified as essentially in translational regulation and cell proliferation. Their dysfunction has been linked to the development of various disease phenotypes and cancers. Using Human BodyMap v2.0 data and data made available through The Cancer Genome Atlas (TCGA), we proposed to observe the patterns of expression of RBPs in sixteen healthy tissues and across nine cancers, and their altering profiles. Additionally, by incorporating BioGrid protein-protein interaction data and CORUM protein complex information, we explore how network properties of the RNA may infer their dysfunction in cancers. The prognostic effect of RBPs classified by expression and network properties in breast cancer were determined using KM-Plotter. We observed that RBPs as a class are more highly expressed than other factors in the 16 human tissues, and furthermore that they are generally upexpressed in cancers. A smaller subset of RBPs (30) is many-fold higher expressed across a large portion of the observed cancers. Network metrics showed no significant differences, except for shortest path distances between subsets (Wilcox, p < 2x10·16). Similarly, complex size and membership did not show any trends or significant differences. The negative prognostic effect seems to be associated with mean path lengths of RBPs and their interaction with a highly dysregulated subset of RBPs.