基于宿主基因表达谱的结核诊断组合标识的发现和同步检测技术研发  

Host-gene expression profiles based discovery and simultaneous detection of gene biosignature for tuberculosis diagnosis

作　　者：郭九标 张惠华 蔡毅 杨倩婷[2] 陈骑[2] 杨帆[1] 张明霞[2] 邓国防[2] 陈心春 GUO Jiu-biao;ZHANG Hui-hua;CAI Yi;YANG Qian-ting;CHEN Qi;YANG Fan;ZHANG Ming-xia;DENG Guo-fang;CHEN Xin-chun(Guangdong Provincial Key Laboratory of Regional Immunity and Diseases,Department of Pathogen Biology,Shenzhen University School of Medicine,Shenzhen,Guangdon 518052;Shenzhen Third People's Hospital,Shenzhen,Guangdong 518112)

机构地区：[1]广东省组织器官区域免疫与疾病重点实验室,深圳大学基础医学院病原生物学系,广东深圳518052 [2]深圳市第三人民医院,广东深圳518112

出　　处：《赣南医学院学报》2021年第8期784-790,共7页JOURNAL OF GANNAN MEDICAL UNIVERSITY

基　　金：广东省组织器官区域免疫与疾病重点实验室项目(2019B030301009);国家自然科学基金青年项目(31800064);深圳市学科布局(重点)项目(JCYJ20180507182049853、JCYJ20200109105012589);广东省绿色发展、可持续发展和人民群众生活需要的科技项目(sgybey01)。

摘　　要：目的:开发结核诊断新标识。方法:利用特征选择方法从20个基因中选取3个基因(CD157、GSDMD和VAMP5)可用于初步区分肺结核患者(59例)与非结核人群(188例);采用Taqman单管同步检测技术验证3个基因的重复性,比较其在新鲜和冻存外周血中的表达量;最后通过神经网络(neural network,NNET)建立用于区分肺结核患者和非结核的分类诊断模型。结果:建立并验证了Taqman单管同步检测3个基因的实时荧光定量技术;通过3个基因的组合(CD157、GSDMD和VAMP5)构建了可用于区分肺结核(158例)与非结核(157例)的NNET模型,该模型对区分肺结核的准确度、灵敏度与特异性分别为0.77(95%CI:0.72,0.82)、0.80(95%CI:0.72,0.86)和0.75(95%CI:0.68,0.82),曲线下面积(area under curve,AUC)为0.85(95%CI:0.80,0.89);在来自独立测试集样本中(包含33个结核样本,35个非结核样本),该模型对区分肺结核的准确度、灵敏度和特异性分别为:0.75(95%CI:0.63,0.85)、0.74(95%CI:0.57,0.88)和0.76(95%CI:0.58,0.89),AUC为0.84(95%CI:0.80,0.89)。结论:鉴定出了能够用于区分和诊断结核的三基因组合标识,开发出了可以单管同步检测分析该3个基因的技术,用于临床上快速筛查诊断结核患者。Objective:To develop novel biomarkers for tuberculosis diagnosis.Methods:Feature selection approach was applied to select three genes from a pool of 20 genes for the differentiation of tuberculosis(59 specimens)from nontuberculosis(188 specimens);modified single-tube probe based RT-qPCR technology was performed to further verify the repeatability and stability of the three-gene biosignature(CD157,GSDMD and VAMP5)in diagnosis tuberculosis from fresh and frozen specimens;a NNET model for efficiently diagnosis of tuberculosis in clinic was established.Results:A NNET model based on the three-gene biosignature(CD157,GSDMD and VAMP5)was built for the diagnosis of tuberculosis(158 specimens)from others(157 specimens),with the accuracy,sensitivity and specificity was 0.77(95%CI:0.72,0.82),0.80(95%CI:0.72,0.86)and 0.75(95%CI:0.68,0.82)respectively,and the AUC of the model was 0.85(95%CI:0.80,0.89).In addition,the NNET model was further verified in an independent cohort of tuberculosis(33 specimens)from others(35 specimens),the resulting accuracy,sensitivity and specificity was 0.75(95%CI:0.63,0.85),0.74(95%CI:0.57,0.88)and 0.76(95%CI:0.58,0.89),with an AUC of 0.84(95%CI:0.80,0.89).Con⁃clusion:A three-gene biosignature(CD157,GSDMD and VAMP5)was identified and a NNET model for the efficiently diagnosis of tuberculosis developed;a single-tube RT-qPCR technology was developed for the convenient clinical diagnosis of tuberculosis from other groups of people.

关 键 词：结核/诊断 神经网络 单管同步检测 

分 类 号：R446.6[医药卫生—诊断学] R52[医药卫生—临床医学]