项目名称: 基于网络解卷积和贝叶斯模型的组学数据融合分析方法及应用研究
项目编号: No.81473072
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 医药、卫生
项目作者: 李康
作者单位: 哈尔滨医科大学
项目金额: 75万元
中文摘要: 通过不同的组学数据,可以在多层次上研究肿瘤的发生与发展机制,融合分析的核心是建立基因组→蛋白组→代谢组之间的关系网络,以及筛选肿瘤标志物。对不同组学数据同时进行分析,研究的最大难点在于不同分子水平上的物质和碎片数量巨大,以及层内和层间的生化物质相互传递和作用关系复杂,在建立调控/互作网络时,变量筛选和网络结构识别极为困难。本研究结合蛋白质组-代谢组同时检测的数据,首先筛选合适的网络变量,然后在不同层面的组学数据上进行特征提取和对接,再使用网络解卷积的方法建立多尺度结点间的直接关联矩阵,最后通过贝叶斯模型确定关联方向。研究的主要内容:网络解卷积的通用算法、关联矩阵的构建和优化、网络解卷积-贝叶斯模型(ND-BN)的融合算法、标志物鉴定及调控网络验证。本课题将在前期代谢组学研究的基础上,给出新的多组学数据融合分析的模型与算法,同时将其应用于卵巢癌标志物的筛选和通路研究上。
中文关键词: 网络解卷积;贝叶斯网络;组学;数据融合分析;卵巢癌标志物
英文摘要: According to the information from different omics data,we could investigate the incidence and mechanism of tumor at different levels. The key point is to establish the correlation network from genomics to proteomics and metabolomics and select the potential biomarkers for carcinoma. When it comes to analyzing the different omics data, one of the most challenging problem is to analyze the large-scale data and mining the internal and external the relationship at different omics data. Therefore, the most important isssue is to select the variables and recognize the most reliable network when we establish the regulation or interactive networks. This study aims to establish a novel data integratio algorithm to establish the reliable network for the tumor diagnosis and mechanism, which combines the information from proteomics and metabolomics to select the potential variables that are used to construct the network. This network is to construct the direct correlation matrix between multiple nodes and then recognize the correlation direction using the Bayesian model. The contents include that network devolution algorithm, constuction and optimization of correlation matrix,integration algorithm and model of network devolution and Bayesian model, biomarker identification and regulation network validation. The project is to propose the novel data integration algorithm and model for multiple omics and apply this algorithm and model to select potential biomarkers and relavent pathway.
英文关键词: network devolution;baysian network;omics;data integration;ovarian cancer biomarker