A Gene Regulatory Network Prediction Method using Particle Swarm Optimization and Genetic Algorithm

Abstract

regulatory networks (GRNs) are complex control systems that deal with the interaction of genes, which eventually control cellular processes at the protein level. The investigation of GRN provides huge information on cellular processes and gene functions and at last contributes to knowledge in genetics and in turn quality of life. By understanding the dynamics of these networks using correct and representative methods and models, potentially cover the way for curing diseases, improving diagnostic procedures and producing drug designs with greater impact. In this work a GRN prediction method based on TDCLR using PSO and GA is proposed to construct GRN from microarray datasets. TDCLR is used to find the directions of information flow between different gene pairs. The proposed method uses the particle swarm optimization (PSO) to find thresholds for discretizing the microarray dataset and genetic algorithm (GA) is used to generate a set of fit candidate gene pair from which GRN is constructed. The sub-network containing five genes of S.cerevisiae (yeast) is used to evaluate the accuracy of the proposed method. The experimental results show that the proposed method is better than TDCLR and other existing methods such as mutual information (MI) in terms of sensitivity and specificity. KeywordsRegulatory Networks (GRN), Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Mutual Information (MI), Context Likelihood of Relatedness (CLR), Time Delay Mutual Information (TDMI), Time Delay Context Likelihood of Relatedness (TDCLR) Many methods are proposed to predict GRNs based on above techniques. In this paper we propose a novel hybrid method which is based on particle swarm optimization (PSO) and genetic algorithm (GA) to predict GRNs. The proposed method also uses TDMI and TDCLR to find the direction between different pairs of genes.