NWE: Node-weighted expansion for protein complex prediction using random walk distances

In this work, we propose a protein complex prediction method based on the method of node-weighted expansion of a cluster, which simulates a random walk with restarts with the cluster whose nodes are weighted. Our implementation of the method is called NWE (Node-Weighted Expansion for protein complex prediction). We have applied NWE to WI-PHI, a genome-wide PPI network of S. cerevisiae, which is obtained by integrating various heterogeneous data sources, and obtained statistically significant clusters of proteins. The biological significance of the predicted clusters of proteins are validated by comparison with manually curated heteromeric protein complexes of S. cerevisiae in the CYC2008 database. In this work, the weight of a node is set to be the total sum of the weights of all the edges adjoining to the node.