基于bootstrap方法的贝叶斯网络结构学习算法在构建基因调控网络中的应用

The Application of Bayes Network Structure Learning Algorithm Based on Bootstrap Method to the Construction of Gene Regulatory Networks

目的探讨基于bootstrap重抽样方法的贝叶斯网络结构学习算法构建网络的性能,并将其应用于卵巢癌基因表达谱数据分析。方法通过模拟实验和实例验证本文给出的算法构建网络的有效性,同时将这种算法应用于构建基因调控网络。结果模拟实验显示,在样本量较小的情况下,基于bootstrap算法构建的贝叶斯网络明显优于普通贝叶斯方法构建的网络;实例分析结果也表明,应用本文的方法能够得到有价值的网络结构。结论应用本文给出的算法能够在样本量较少的情况下得出准确度较高的网络,同时能够给出网络结构中各条边置信度的估计值。

Objective To explore the performance of Bayes network structure learning algorithm based on bootstrap method in network construction, and to apply it to the analysis of ovarian cancer gene expression data. Methods The efficiency of the algorithm given in this article was testified with simulation data and gene expression data, and meanwhile this algorithm was used to construct gene regulatory networks. Results Bayes network structure learning based on bootstrap method performed better than the general Bayes network in the case of small sample sizes, as shown in simulation tests; the results of gene expres- sion data analysis also indicated that this algorithm could provide valuable network structures. Conclusion Bayes network struc- ture learning algorithm based on bootstrap method can establish highly precise network models even with small sample sizes, and meanwhile provide the confidence estimates of each edge in the network.