Li et al. (2010) constructed a relatively reliable weighted PPI network by taking advantage of gene annotations. In this weighted network, the performance of the topology-based methods could be optimized to a great extent. This result confirms that topological centrality is sensitive to the noise in PPI networks. To further improve the prediction precision, researchers attempt to combine biology information with network topology (He and Zhang, 2006; Zotenko et al., 2008; Chua et al., 2008).