Title

Reconstruction of genetic network by Bayesian network model with integration of various prior knowledge

Date of Completion

January 2010

Keywords

Biology, Systematic|Biology, Bioinformatics|Computer Science

Degree

Ph.D.

Abstract

Bayesian network model is widely used for reverse engineering of gene regulatory network structure. An advantage of this model lies in its capability to integrate prior knowledge into the model learning process, which can lead to improvement in the quality of the analysis outcome. Sonic previous works have explored this area. Unfortunately, most of the existing works focus only on prior knowledge of the direct, variable links. Here we propose a set of methods designed to integrate other types of prior knowledge in model learning, namely, the semantic variable relations and indirect variable relations. We show in this work how these knowledge can be formalized and integrated into the model definition, and how the resulting models are evaluated with simulated data and real biological data. It has been shown that the integration of prior knowledge results in a significant improvement in their model performances. To address the issue of prior knowledge generation, we also proposed an approach to automatically extract indirect variable links from knowledge databases such KEGG and GO. ^