Learning Gene Regulatory Network from Microarrays Based on Bayesian Network

Abstract

The development of modern medical technology gives a big boost to analyze microarrays and construct gene regulatory network. To get a clean and straightforward gene regulatory network from enormous microarray expression data, we propose a learning algorithm based on Bayesian network. We use K-Medoids to obtain medoid genes and greedy search with BDe to construct Bayesian network structure related to the certain disease. Experimental result on REGED0 demonstrates that the gene regulatory network of lung cancer we get from microarrays contains representative genes and reveals core regulatory relationships with high effectiveness and efficiency.