基因调控网络的父节点筛选贝叶斯建模方法

在构建基因调控网络的方法中，贝叶斯网络模型可以直观地表达基因间的调控关系，但在结构学习时的复杂度极高，使得网络建模效率较低且规模有限.因此，本文提出一种基于父节点筛选的贝叶斯网络(parent node screening based Bayesian network， PS-BN)建模方法.PS-BN方法将关联模型与贝叶斯网络模型相结合，在充分利用贝叶斯网络模型结构学习搜索策略的前提下，先基于父节点筛选方法去除部分冗余信息，以达到缩减搜索空间的目的.实验结果表明，与传统的贝叶斯网络模型方法相比，PS-BN方法极大提升了基因调控网络构建效率，同时准确率有所提高.

Modelling of Gene Regulatory Networks by Parent Node Screening-Based Bayesian Method

Among the methods for modeling gene regulation networks， Bayesian network model can intuitively express the regulatory relationship between genes. However， due to the high complexity of Bayesian network model in the structure learning， the efficiency of the gene regulation networks modeling is low and the scale of the reconstructed network is limited. Therefore， this paper proposed a method which is called the parent node screening-based Bayesian network(PS-BN). The PS-BN method combines the correlation model with Bayesian network model. Under the premise of making full use of the search strategy of structure learning in Bayesian network model， the parent node screening method is used to remove some redundant nodes， thus reducing the search space. The experimental results show that compared with the Bayesian network model， the PS-BN method greatly improves the efficiency of modeling gene regulatory networks while improving the accuracy.