Inferring gene regulatory networks from multiple data sources via a dynamic bayesian network with structural EM

Authors:
Yu Zhang;Zhidong Deng;Hongshan Jiang;Peifa Jia
Affiliations:
State Key Laboratory of Intelligent Technology and System, Computer Science and Technology Department, Tsinghua University, Beijing, China;State Key Laboratory of Intelligent Technology and System, Computer Science and Technology Department, Tsinghua University, Beijing, China;Department of Computer Science, Tsinghua University, Beijing, China;State Key Laboratory of Intelligent Technology and System, Computer Science and Technology Department, Tsinghua University, Beijing, China
Venue:
DILS'07 Proceedings of the 4th international conference on Data integration in the life sciences
Year:
2007

Citing 3
Cited 3

Advances to Bayesian network inference for generating causal networks from observational biological data

Bioinformatics
Predicting the prognosis of breast cancer by integrating clinical and microarray data with Bayesian networks

Bioinformatics
Learning the structure of dynamic probabilistic networks

UAI'98 Proceedings of the Fourteenth conference on Uncertainty in artificial intelligence

An approach for RNA secondary structure prediction based on Bayesian network

CIBCB'09 Proceedings of the 6th Annual IEEE conference on Computational Intelligence in Bioinformatics and Computational Biology
Using Qualitative Probability in Reverse-Engineering Gene Regulatory Networks

IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Inferring Gene Regulatory Networks via Nonlinear State-Space Models and Exploiting Sparsity

IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)

Quantified Score

Hi-index	0.01

Visualization

Abstract

Using our dynamic Bayesian network with structural Expectation Maximization (SEM-DBN), we develop a new framework to model gene regulatory network from both gene expression data and transcriptional factor binding site data. Only based on mRNA expression data, it is not enough to accurately estimate a gene network. It is difficult for us to estimate a gene network accurately only with the mRNA expression data. In this paper, we use the transcription factor binding location data in order to introduce the prior knowledge to SEM-DBN model. Gene expression data are also exploited specifically for likelihood. Meanwhile, we incorporate the prior knowledge into every learning step by SEM rather than only learning from the very beginning, which can compensate the attenuation of the effect with location data. The effectiveness of our proposed method is demonstrated through the analysis of Saccharomyces cerevisiae cell cycle data. The combination of heterogeneous data from multiple sources ensures that our results are more accurate than those recovered from only gene expression data alone.