基于功能网络的疾病microRNA预测及模块识别方法

MicroRNA (miRNA) function prediction is one of popular issues in bioinformatics. The project will develop the methods of constructing a functional network composed of miRNAs. On the basis of the functional network, it will also investigate the methods of predicting disease-related microRNAs and identifying of their functional module to explore their function in the disease progression. First, a weighted Bayes method is proposed to create a reliable target dataset, which completely considers the characteristics of multiple target prediction methods. Second, we present a miRNA functional similarity measurement which combining interactive context of targets. Furthermore, the threshold selection based on clustering coefficient is used to construct the miRNA functional network. Third, the bilayer network is constructed by integrating the multiple information, such as the protein-protein interactions and the functional similarity between miRNAs. Moreover, the disease-related miRNA prediction method is proposed based on the bilayer network. The method can predict the miRNAs associated with a given disease. Finally, we investigate the functional module identification method based on propagating multiple disease labels. The propagation strength is given to describe the consistence of two groups of diseases associated with miRNAs. The method can be used to identify the miRNA functional module related to a specific disease classification. The project has great significance for exploiting the new computational prediction methods of disease-related miRNAs. It also has potential application value in exploring mechanisms of human disease.

MicroRNA（miRNA）的功能预测是当前生物信息学领域的研究热点。本项目首先发展由miRNA结点组成的功能网络构建方法，并在此基础上研究疾病miRNA预测及其功能模块的识别方法，以探索其在疾病中行使的功能。首先，提出一种考虑多种靶基因预测方法特点的贝叶斯加权方法，以建立可靠的靶基因集合。其次，提出一种结合靶基因互作上下文的miRNA功能相似性计算方法，进而基于聚类系数的阈值选择方法构建功能网络。然后，整合蛋白互作关系、miRNA间功能相似等多种信息建立双层网络，并提出基于双层网络的疾病miRNA预测方法，有效预测与给定疾病关联的miRNA。最后，研究基于疾病标签传播的模块识别方法，提出“传播力度”用以描述miRNA间关联疾病的一致性，有效识别疾病类别相关的miRNA功能模块。本项目对开拓新的疾病miRNA信息学预测方法有重要的理论意义，并在探索人类疾病发生机理方面具有潜在的应用价值。

MicroRNA(miRNA)的功能预测是当前生物信息学领域的研究热点。本项目首先发展了由miRNA结点组成的功能网络构建方法，并在此基础上研究了疾病miRNA预测及其功能模块的识别方法，以探索其在疾病中行使的功能。首先，提出了一种考虑多种靶基因预测方法特点的贝叶斯加权方法，建立了可靠的靶基因集合。其次，提出了一种结合靶基因互作上下文的miRNA功能相似性计算方法，进而基于聚类系数的阈值选择方法构建功能网络。然后，整合了蛋白互作关系、miRNA间功能相似等多种信息建立双层网络，并提出基于双层网络的疾病miRNA预测方法，有效地预测了与给定疾病关联的miRNA。最后，尝试研究了基于疾病标签传播的模块识别方法，提出了“传播力度”用以描述miRNA间关联疾病的一致性，有效识别了疾病类别相关的miRNA功能模块。本项目对开拓新的疾病miRNA信息学预测方法有重要的理论意义，并在探索人类疾病发生机理方面具有潜在的应用价值。