肉碱棕榈酰基转移酶2与嗅介蛋白样3是诊断颈动脉粥样硬化不稳定斑块的分子标志物  

Carnitine palmitoyltransferase Ⅱ and olfactomedin like 3 are molecular markers for the diagnosis of unstable carotid atherosclerotic plaques

作　　者：化召辉[1] 周保宁 张帅[1] 张林枫[1] 马珂[1] 李震[1] Hua Zhaohui;Zhou Baoning;Zhang Shuai;Zhang Linfeng;Ma Ke;Li Zhen(Department of Endovascular Surgery,the First Affiliated Hospital of Zhengzhou University,Zhengzhou 450052,China)

机构地区：[1]郑州大学第一附属医院腔内血管外科,郑州450052

出　　处：《中华实验外科杂志》2023年第1期131-133,共3页Chinese Journal of Experimental Surgery

基　　金：河南省高等学校重点科研项目计划(22A320081)。

摘　　要：目的探讨不稳定型颈动脉粥样硬化斑块的分子标志物并分析其与免疫浸润的相关性。方法通过基因表达综合数据库(GEO)下载颈动脉斑块RNA芯片数据(GSE43292),共纳入64例颈动脉斑块样本(稳定型:32例;不稳定型:32例)。将64例颈动脉斑块(稳定型:32例;不稳定型:32例)的全转录组数据进行加权基因共表达网络分析,采用一致性聚类识别离群样本并运用动态剪切树算法选择网络中的关键模块。应用随机森林、支持向量机与Lasso回归等多种机器学习算法分析模块中的基因与颈动脉粥样硬化不稳定斑块的相关性,采用ROC曲线评价关键基因对不稳定型斑块的预测价值。结果(1)每例样本均得到23307个基因。加权基因共表达网络分析显示当软阈值为16时符合无尺度网络,2例样本为离群样本,黄色模块与不稳定斑块相关性最高为0.6,共334个基因。(2)分析334个基因与不稳定斑块的相关性:随机森林算法共得到CMPK1、CTP2、KLRD1、PAK1等89个关键基因;Lasso回归算法共得到嗅介蛋白样3(olfactomedin like 3,OLFML3)、DTX3L、DTX3L、SSH2等8个关键基因;支持向量机算法共得到CTP2、OLFML3、NCF2、FHAD1等11个关键基因。(3)肉碱棕榈酰基转移酶2(CPT2)与OLFML3为3种算法的共同关键基因,ROC曲线分析显示,CPT2、OLFML3预测颈动脉粥样硬化不稳定斑块的曲线下面积分别为0.82[95%可信区间(CI):0.71~0.93],0.77(95%CI:0.65~0.89);CPT2与OLFML3共同预测颈动脉粥样硬化不稳定斑块的曲线下面积为0.82(95%CI:0.72~0.92)。结论CPT2与OLFML3对颈动脉粥样硬化不稳定斑块具有良好的预测价值。Objective To investigate the molecular markers of unstable carotid atherosclerotic plaque and analyze the correlation between unstable carotid atherosclerotic plaque and immune infiltration.Methods Carotid plaque RNA chip data(GSE43292)were downloaded from the Gene Expression Omnibus(GEO)database,and 64 carotid plaque samples were included(stable type:32 cases;Unstable type:32 cases).The whole transcriptome data of 64 cases of carotid plaques(stable type:32 cases;unstable type:32 cases)were analyzed by weighted gene co-expression network.Consistent clustering was used to identify outlier samples and dynamic cut tree algorithm was used to select key modules in the network.Random forest,support vector machine,Lasso regression and other machine learning algorithms were used to analyze the correlation between genes in the module and unstable carotid atherosclerotic plaque,and ROC curve was used to evaluate the predictive value of key genes for unstable plaque.Results(1)A total of 23307 genes were obtained from each sample.Weighted gene coexpression network analysis showed that when the soft threshold was 16,it accorded with the scale-free network.There were 2 outlier samples,and the correlation between yellow module and unstable plaque was the highest(334 genes).(2)Tthe correlation between 334 genes and unstable plaque was analyzed:89 key genes such as CMPK1,CTP2,KLRD1,PAK1 were obtained by random forest algorithm,8 key genes such as olfactomedin like 3(OLFML3),DTX3L,DTX3L,SSH2 were obtained by Lasso regression algorithm,and 11 key genes such as CTP2,OLFML3,NCF2,FHAD1 were obtained by support vector machine algorithm.(3)Carnitine palmitoyltransferase Ⅱ(CPT2)and OLFML3 were the common key genes of the three algorithms.ROC curve analysis showed that the areas under the curve for predicting unstable carotid atherosclerotic plaques by CPT2 and OLFML3 were 0.82[95% confidence interval(CI):0.71-0.93]and 0.77(95%CI:0.65-0.89),respectively,and the areas under the curve predicted by CPT2 and OLFML3 were 0.82(95%CI:0.72-0.92

关 键 词：颈动脉 不稳定斑块 生物标志物 机器学习 

分 类 号：R543.4[医药卫生—心血管疾病]