LSPC: An Algorithm for Inference of Gene Networks Using Bayesian Network

Gene regulatory networks explain how cells control the expression of genes, which, together with some additional regulation downstream, determines the production of proteins essential for cellular function. Bayesian networks (BNs) are practical tools which have been successfully implemented in learning gene networks based on microarray gene expression data. Bayesian networks are graphical representation for probabilistic relationships among a set of random variables. PC algorithm is a structure learning algorithm based on conditional independence tests. The drawback of PC algorithm is that high-order conditional independence (CI) tests need large sample sizes. The number of records in microarray dataset is rarely enough to perform reliable high-order CI tests. In this paper, we extend the methodology for reduction of the order of the CI tests. In order to improve the PC algorithm, we introduce a heuristic algorithm, LSPC, for learning the structure of the BN. The results indicate that applying the LSPC methodology improves the precision of learning the skeleton of the graph (undirected graph) for Bayesian networks