Cheng Ming, An Shoukuan, Li Junquan
Department of Cardiac Surgery, The Second Affiliated Hospital of Harbin Medical University, Nangang, Harbin 150086, Heilongjiang, People's Republic of China.
Department of Cardiac Surgery, The Second Affiliated Hospital of Harbin Medical University, Nangang, Harbin 150086, Heilongjiang, People's Republic of China..
Gene. 2017 Sep 20;629:101-107. doi: 10.1016/j.gene.2017.07.070. Epub 2017 Jul 29.
Coronary artery disease (CAD) has a high mortality rate and consists of multiple condition, including stable/unstable angina, sudden cardiac death, and myocardial infarction. This study is aimed to explore the pathogenesis of CAD.
Datasets of GSE20680 (including 87 CAD samples and 52 normal samples) and GSE20681 (including 99 CAD samples and 99 normal samples) were obtained from Gene Expression Omnibus database. The differentially expressed genes (DEGs) were identified by MetaDE. Effect Sizes in MetaDE package, and then were hierarchical clustered using pheatmap package in R. Subsequently, CAD-associated microRNAs (miRNAs) and their targets were obtained separately by miR2Disease and miRTarBase databases, and then used to construct an associated-miRNA-DEG regulatory network based on BioGRID, HPRD and DIP databases. Enrichment analysis was conducted for the involved DEGs using Fisher's exact test, and a support vector machine (SVM) classifier was constructed to optimize the feature genes. After CAD-associated long non-coding RNAs (lncRNAs) were predicted by lncRNA Disease database and their target miRNAs were predicted using miRcode and starBase databases, lncRNA-miRNA-DEG regulatory network was constructed.
Total 1208 DEGs were screened, and 5 CAD-associated miRNAs (including miR-92a) were predicted associated with CAD. The SVM classifier was constructed based on the 41 featured genes and had high recognition efficiency. Only one lncRNA CDKN2B-AS targeting miR-92a was obtained. Finally, GATA2, MAP1B and ARG1 were involved in the CDKN2B-AS-miR-92a-feature gene regulatory network.
GATA2, MAP1B and ARG1 indirectly regulated by CDKN2B-AS through miR-92a might be involved in CAD.
冠状动脉疾病（CAD）死亡率高，由多种病症组成，包括稳定型/不稳定型心绞痛、心源性猝死和心肌梗死。本研究旨在探索CAD的发病机制。
从基因表达综合数据库获取GSE20680数据集（包括87个CAD样本和52个正常样本）和GSE20681数据集（包括99个CAD样本和99个正常样本）。通过MetaDE识别差异表达基因（DEG）。使用MetaDE软件包中的效应大小，然后在R语言中使用pheatmap软件包进行层次聚类。随后，分别通过miR2Disease和miRTarBase数据库获得CAD相关的微小RNA（miRNA）及其靶标，然后基于BioGRID、HPRD和DIP数据库构建相关miRNA-DEG调控网络。使用Fisher精确检验对涉及的DEG进行富集分析，并构建支持向量机（SVM）分类器以优化特征基因。通过lncRNA疾病数据库预测CAD相关的长链非编码RNA（lncRNA），并使用miRcode和starBase数据库预测其靶标miRNA，构建lncRNA-miRNA-DEG调控网络。
共筛选出1208个DEG，预测有5个与CAD相关的miRNA（包括miR-92a）与CAD相关。基于41个特征基因构建了SVM分类器，具有较高的识别效率。仅获得一个靶向miR-92a的lncRNA CDKN2B-AS。最后，GATA2、MAP1B和ARG1参与了CDKN2B-AS-miR-92a-特征基因调控网络。
CDKN2B-AS通过miR-92a间接调控的GATA2、MAP1B和ARG1可能参与CAD。
