View article

Protein-Protein Interaction (PPI) networks are believed to be important sources of information related to biological processes and complex metabolic functions of the cell. The presence of biologically relevant functional modules in PPI networks has been theorized by many researchers. However, PPI networks are known to contain noisy false positive interactions and possess the scale-free property, which makes the task of isolating these useful modules difficult. In this paper, we propose an ensemble clustering approach to address this problem. To perform initial clustering, we examine three topology-based distance metrics that are conducive for partitioning these networks. To perform consensus clustering, we develop a PCA-based hypergraph approach, designed to handle large interaction networks. We also develop a soft consensus clustering method to assign multifaceted hub proteins to multiple functional …