Graph Theory Analysis of Protein-Protein Interaction Graphs Through Clustering Method

Graph mining is an active progressive field of research in recent years to infer knowledge from complex data. Graphs can model the data in an efficient manner especially when the data is relation specific. One of the graph mining technique is graph clustering which is used to discover groups or clusters of nodes that share common characteristics. The function of a cell is performed by interacting proteins. Each protein is responsible for one or more function. Pair wise interaction of proteins make a protein complex. And each protein complex performs a specific function. Finding protein complexes helps to detect function of unknown proteins in a way that proteins belonging to a complex have same function. Protein-protein interaction data can be cast as a graph such that vertices represent proteins and edges represent interaction between proteins. Graph based clustering method can be then applied to detect the set of connected proteins known as protein complexes. Identification of protein complexes related to diseases may be considered as an important aspect for drug design and curation. Louvain cluster is an uncomplicated and effective implementation of clustering method to detect dense clusters.