基于机器学习的体外循环后心肌缺血-再灌注损伤中m6A相关基因聚类分析和免疫浸润分析  

m6A-related gene clustering analysis and immune cell infiltration analysis in myocardial ischemia-reperfusion injury after cardiopulmonary bypass based on machine learning

作　　者：唐垚 陈文栋[1] 王燕琼 杨伟[1] TANG Yao;CHEN Wendong;WANG Yanqiong;YANG Wei(Department of Anesthesiology,The First Affiliated Hospital of Kunming Medical University,Kunming,650032,P.R.China)

机构地区：[1]昆明医科大学第一附属医院麻醉科,昆明650032

出　　处：《中国胸心血管外科临床杂志》2024年第10期1475-1485,共11页Chinese Journal of Clinical Thoracic and Cardiovascular Surgery

基　　金：云南省科技厅-昆明医科大学应用基础研究联合专项基金(2019FE001[-045]);云南省后备人才基金(H-2019028)。

摘　　要：目的通过机器学习的方法确定体外循环后心肌缺血-再灌注损伤(myocardial ischemia-reperfusion injury,MI/RI)中N6-甲基腺苷(N6-methyladenosine,m6A)相关的特征基因并对其进行聚类分析及免疫浸润分析。方法通过GEO中数据集GSE132176,筛选m6A甲基化相关的差异基因,基于差异基因表达谱对该数据集样本进行聚类分析。并对聚类后m6A簇的差异基因进行基因本体论(GO)和京都基因与基因组百科全书(KEGG)富集分析以确定m6A簇基因功能。通过R软件从支持向量机(SVM)模型和随机森林(RF)模型中确定更优模型,用于筛选MI/RI中m6A相关的特征基因,并构建特征基因列线图以预测疾病的发病率。运用R软件分析特征基因与免疫细胞的相关性,并运用在线网站构建特征基因调控网络。结果该数据集中共筛选出5个m6A相关的差异基因,对其进行聚类分析将基因表达谱分为两簇。对m6A簇进行富集分析,结果显示这些基因主要参与调节单核细胞分化、对脂多糖的反应、对细菌来源分子的反应、细胞对氧气水平下降的反应、DNA转录因子结合、DNA结合转录激活剂活性、RNA聚合酶Ⅱ特异性、NOD样受体信号通路、流体剪切应力与动脉粥样硬化、肿瘤坏死因子(TNF)信号通路、白介素(IL)-17信号通路等。利用R软件确定RF模型为更优模型,该模型确定METTL3、YTHDF1、RBM15B、METTL14为MI/RI的特征基因,免疫浸润分析发现肥大细胞、1型辅助性T淋巴细胞(Th1)、17型辅助性T淋巴细胞(Th17)、巨噬细胞与体外循环后MI/RI相关。结论机器学习获得4个特征基因METTL3、YTHDF1、RBM15B、METTL14,同时聚类分析和免疫细胞浸润分析更好地揭示体外循环后MI/RI可能存在的病理生理过程。Objective To identify the N6-methyladenosine(m6A)-related characteristic genes analyzed by gene clustering and immune cell infiltration in myocardial ischemia-reperfusion injury(MI/RI)after cardiopulmonary bypass through machine learning.Methods The differential genes associated with m6A methylation were screened by the dataset GSE132176 in GEO,the samples of the dataset were clustered based on the differential gene expression profile,and the Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)enrichment analysis of the differential genes of the m6A cluster after clustering were performed to determine the gene function of the m6A cluster.R software was used to determine the better models in machine learning of support vector machine(SVM)model and random forest(RF)model,which were used to screen m6A-related characteristic genes in MI/RI,and construct characteristic gene nomogram to predict the incidence of disease.R software was used to analyze the correlation between characteristic genes and immune cells,and the online website was used to build a characteristic gene regulatory network.Results In this dataset,a total of 5 m6A-related differential genes were screened,and the gene expression profiles were divided into two clusters for cluster analysis.The enrichment analysis of m6A clusters showed that these genes were mainly involved in regulating monocytes differentiation,response to lipopolysaccharides,response to bacteria-derived molecules,cellular response to decreased oxygen levels,DNA transcription factor binding,DNA-binding transcription activator activity,RNA polymeraseⅡspecificity,NOD-like receptor signaling pathway,fluid shear stress and atherosclerosis,tumor necrosis factor signaling pathway,interleukin-17 signaling pathway.The RF model was determined by R software as the better model,which determined that METTL3,YTHDF1,RBM15B and METTL14 were characteristic genes of MI/RI,and mast cells,type 1 helper lymphocytes(Th1),type 17 helper lymphocytes(Th17),and macrophages were found to be assoc

关 键 词：体外循环 N6-甲基腺苷RNA甲基化 心肌缺血-再灌注损伤 机器学习 

分 类 号：R654[医药卫生—外科学]