A Novel Method for Inference of Genetic Regulatory Networks Using Time Series Microarray Data and Gene Ontology

Bayesian Network (BN) models have been used in reconstructing Gene Regulatory Networks (GRNs) from gene expression microarray data for their advantages. However,using only microarray data is not sufficient for accurate estimating a GRN. Here, a novel two step method is proposedfor reconstructing GRNs using both time series microarray dataand Gene Ontology (GO): first, genes are divided into some groups based on GO annotations and then BN is applied toinfer the gene relations in each group. In the latter step, a combination of GO and microarray data is used to evaluatedifferent network structures and find the optimal one. This method is applied to reconstruct the regulatory network of 84 yeast genes. Comparing the simulation results with the KEGGpathway map shows that the method improves the accuracy of the estimated gene network from 62% to 73% and the sensitivity by increasing the number of True Positives from 55 to 82.

Keywords:

Bayesian Network , BN , Gene Regulatory Network , GRN , Gene Ontology , GO , Time Series Microarray Data ,