Reverse Engineering of Gene Regulatory Network by Integration of Prior Global Gene Regulatory Information

A Bayesian network is a model to study the structures of gene regulatory networks. It has the ability to integrate information from both prior knowledge and experimental data. Some previous works have explored the advantage of using prior knowledge. Unfortunately, most of the existing works only utilize biological knowledge about local structures of each gene in the network. In this study, we propose an approach to efficiently integrate global ordering information into model learning, where the ordering information specifies the indirect relationships among genes. We study the model behaviors with synthetic data. We demonstrate that, compared with a traditional Bayesian network model that uses only local prior knowledge, utilizing additional global ordering knowledge can significantly improve the modelpsilas performance. The magnitude of this improvement depends on how much global ordering information is integrated and how much noise the data includes.