Reverse engineering module networks by PSO-RNN hybrid modeling

Finding a transcriptional regulatory network (TRN) is usually an under-determined problem. To address this challenge, we have developed a novel TRN inference method by integrating gene expression data and gene functional category information. The inference is based on module network model. A module is a set of genes with similar expression profiles, and a network represents regulatory relationships between the modules. The proposed method consists of two parts: the module selection part determines the modules with fuzzy c-mean (FCM) clustering by incorporating gene functional category information, and the network inference part uses a hybrid of particle swarm optimization and recurrent neural network (PSO-RNN) methods to infer the underlying network between modules. Our method was tested on real data from two studies: the development of rat central nervous system and the yeast cell cycle process. The results were validated with comparison to various literature sources and gene ontology biological process information.