Inference of Large-Scale Gene Regulatory Networks Using GA-Based Bayesian Network and Biological Knowledge

A fundamental issue in understanding the biological cellular behavior is based on discovering the interactions between genes, which is known as the gene regulatory network. This paper proposes a novel method to model large-scale gene regulatory networks from time series gene expression data. In the first step, a novel Gene Ontology (GO)-based clustering algorithm is applied to classify genes into smaller sets. In the next step, a combination of Genetic Algorithm (GA) and Bayesian Network (BN) is utilized to model causal relationships between genes in each cluster. In order to improve the search, in addition to microarray data, Protein-Protein Interactions are utilized. We have tested our method on 98 yeast genes from cell cycle gene expression data set collected by Spellman. In comparison to KEGG pathway map, this method is capable of finding 45.66% of true interactions between genes.