Analyzing Rhabdomyosarcoma using the Multimodal Clustering Approach (DReiM)

Abstract

Rhabdomyosarcoma is a type of cancer that is connected to soft tissue, connective tissue, or bone. Every year 350 children are diagnosed with Rhabdomyosarcoma. Majority of the kids diagnosed with this disease are under ten years of age. Though the intensive conventional approach to treatment exists, patients are still at high risks to this aggressive disease. The need to gain understanding and insight into this disease can help the design of therapeutic agents. We utilized a multimodal network approach to gain an understanding of this mechanism. Protein phosphorylation has been mostly studied as a post-translational modification in eukaryotes. They play a significant role in various cellular processes. The mechanism of its oncogenes is not well known with various levels of signaling protein dysregulation. In the Phosphosite database, some proteins phosphorylate to cause Rhabdomyosarcoma. Our method utilizes a co-clustering approach using multimodal networks to analyze the rhabdomyosarcoma network. Rhabdomyosarcoma network in general consists of several heterogeneous networks that include gene-pathway, pathway-drug, and gene-drug. We reconstruct the phosphorylation network for Rhabdomyosarcoma, by creating a network that consists of different types of nodes. The goal is to implement this clustering approach to identify a potential candidate for Rhabdomyosarcoma. We applied network centrality measures to find the most influential nodes first and foremost and then used the clustering approach stated above towards drug repositioning.