随着全球人口老龄化的加快，阿尔茨海默病已经成为严重威胁老年人身体健康的疾病。本项目将利用当前丰富的阿尔茨海默病高通量测序数据，通过构建生物网络并结合miRNA臂转换，将疾病相关的非编码RNA串联起来，应用先进的机器学习方法挖掘更具生物意义的非编码RNA生物标志物。本项目拟开展如下研究：1.构建共表达网络，利用多信息融合方法预测miRNA应答元件并构建ceRNA网络，最后整合为CO-CE网络；2.研发针对CO-CE网络的网络嵌入与改进的AP聚类方法，以进行网络功能模块挖掘；3.利用miRNA-Seq数据进行miRNA臂转换的挖掘，结合CO-CE网络与功能富集分析，对miRNA臂转换在发病过程中的作用进行机理分析；4.利用机器学习方法对网络功能模块和臂转换miRNA进行筛选，以获得高灵敏度和高特异度的生物标志物。研究成果将为阿尔茨海默病的早期诊断和相应的药物研发提供新颖的方法和相关的技术信息。

As the acceleration of global population aging, Alzheimer's disease (AD) has become a serious threat to the health of the seniors. In this study, the current abundant high-throughput sequencing data of AD will be used to build a biological network combining with miRNA arm switching to connect AD-related non-coding RNAs (ncRNAs), and then apply cutting edge machine learning methods to mine more biological significance ncRNA biomarkers. This proposal includes the following study: 1. Construct a co-expression network, and construct a ceRNA network with a multi-information fusion method which can predict miRNA response elements (MREs). Then merge the co-expression network and ceRNA network into the CO-CE network; 2. Develop a network embedding and improved AP clustering method for comprehensive mining of functional modules in CO-CE network; 3. Use miRNA-Seq data to perform miRNA arm switching mining, combining with CO-CE network and gene set enrichment analysis, to analyze the mechanism of miRNA arm switching in the pathogenesis of AD; 4. Apply machine learning methods to screen functional modules and arm switched miRNAs to obtain AD biomarkers with high sensitivity and specificity. The outcome of this project will provide necessary technical information for the early diagnosis of AD and the corresponding drug development.