Clustering of protein-protein interaction network using fractal dimension of protein subnetworks

Abstract

Protein-protein interactions play a vital role in the biological processes of all living organisms. These interactions can be represented as networks, in which a node represents a protein and an edge represents an interaction between a pair of proteins. Clustering of these networks leads to the detection of significant protein complexes. FDPClus, a density based clustering method of these networks using the principles of fractal dimension is proposed here. A modified sand box algorithm is used to find the fractal dimension of protein subnetworks. The F-measure values obtained for the Gavin and Collins data set are 0.48 and 0.63 respectively when compared against the CYC2008 yeast benchmark protein complex set. The proposed method shows better performance than other existing methods such as DPClus, MCODE, RNSC, CORE and MCL. Hence it demonstrates the usefulness of fractal dimension of protein subnetworks in the clustering of Protein-Protein Interaction (PPI) networks.