A method for identifying functional modules from the functionally-associated, significantly-co-expressed, and reliably-linked temporal dynamic PPI networks

A protein functional module is an ensemble of proteins participating in a specific biological process. Proteins in a functional module temporally and dynamically interact to perform their biological function. According to biological behaviors of proteins, identifying functional modules becomes one of major tasks in system biology. In our work, based on protein-protein interaction data, we firstly integrate function annotations, gene expression data, and interaction reliability score to construct functionally-associated, significantly-co-expressed, reliably-linked temporal dynamic protein-protein interaction networks FER-TDPINs. Subsequently, we propose a novel functional module identification method IFM-FER-TDPINs, which exploit seed-expanding strategy, follow a criterion of expanding locally dense connection, cluster the multi-network distributed and function-associated proteins to identify functional modules with expression relevance inside module higher than outward expression relevance. Finally, we test the performances of our proposed method in comparison with other six existing methods on FER-TDPINs constructed by combining three PPI data sets and two gene expression data sets. The experimental results show that our proposed method IFM-FER-TDPINs accurately identified a large number of functional modules and obtained significantly BP term-enriched functional modules with statistically higher qualityWe here provide IFM-FER-TDPINs software suit and the related dataset files.