基于单细胞测序探讨中性粒细胞相关基因在心梗发生中的作用  

Screening of neutrophil-related genes associated with myocardial infarction based on machine learning methods and single cell RNA sequencing

作　　者：唐婕琼 沈童童[1] TANG Jieqiong;SHEN Tongtong(Department of Cardiology,Chuzhou Hospital Affiliated to Anhui Medical University,Chuzhou,Anhui 239000,China)

机构地区：[1]安徽医科大学附属滁州医院(安徽医科大学第五临床医学院)心血管内科,安徽滁州239000

出　　处：《淮海医药》2025年第2期111-116,共6页Journal of Huaihai Medicine

基　　金：2022年滁州市卫生健康科研项目(2022009)。

摘　　要：目的:利用机器学习和单细胞测序技术筛选出心肌梗死(MI)相关的中性粒细胞基因。方法:从基因表达数据库(GEO)中获取MI相关数据集GSE66360、GSE48060和单细胞数据集GSE163465。利用Seurat包筛选出GSE163465中的差异表达基因(DEGs),利用WGCNA包筛选出GSE66360中与中性粒细胞和MI相关的模块基因,得到共同DEGs。通过最小绝对收缩和选择算法(LASSO)、支持向量机递归特征消除(SVM-RFE)和随机森林算法(RF)筛选出关键基因,在GSE48060中验证基因的预测价值。选取预测价值高的基因构建诊断模型,验证模型的诊断效能。结果:通过分析GSE66360和GSE163465数据集,得到50个共同DEGs,功能富集分析表明DEGs与炎症和中性粒细胞相关通路有关,通过机器学习得到4个预测价值高的基因,构建的诊断模型的预测价值为0.913。结论:中性粒细胞相关基因BST1、CLEC4D、SIRPA、TLR2与急性MI的发生有关。Objective:To screen neutrophil-related genes associated to myocardial infarction(MI)by machine learning methods and single cell sequencing.Methods:MI related datasets GSE66360,GSE48060 and GSE163465 from Gene Expression Omnibus(GEO)were obtained.The Seurat package in R language was used to screen differentially expressed genes(DEGs)in dataset GSE163465 and the WGCNA package was employed to identify the key module related to neutrophil and MI based on GSE66360 so as to gain common DEGs.The hub genes were identified by least absolute shrinkage and selection operator(LASSO),support vector machine-recursive feature elimination(SVM-RFE)and random forest(RF)methods.The hub genes were validated in GSE48060 and the genes with AUC>0.8 were integrated to construct a diagnostic model.Then the diagnostic efficiency of the model was versified by AUC curve.Results:50 common DEGs were screened through analyzing datasets GSE163465 and GSE66360.GO analysis displayed that the inflammation-and neutrophil-related pathways were associated with the common DEGs.Four genes with AUC>0.8 were identified by machine learning methods and the AUC of the diagnostic model was 0.913.Conclusion:Four neutrophil-related genes BST1,CLEC4D,SIRPA and TLR2 are related to the occurrence of MI.

关 键 词：心肌梗死 单细胞测序 机器学习 生物信息学 差异表达基因 

分 类 号：R542.22[医药卫生—心血管疾病]