摘要
目的评价基于非参数正态图模型的差异网络分析方法筛选差异基因的效果,并筛选影响肝细胞癌患者生存时间的枢纽基因,用于预测患者的预后或作为新的药物靶点。方法模拟实验中用AUC值评价基于非参数正态图模型的差异网络分析筛选差异基因的效果,再从TCGA数据库下载3个平台5个数据集的肝细胞癌患者的基因表达数据,纳入通路等先验信息利用非参数正态图模型构建差异表达网络,按照节点连接边的个数即度的大小选择枢纽基因。结果在4种条件的模拟实验中,非参数正态图模型筛选差异基因的AUC值范围为0.71～0.97;在肝细胞癌数据的分析中,筛选出8个与肝细胞癌生存时间相关的枢纽基因,分别是IGF1、ErbB2、FGF9、GH2、cSH2、HSP90AA1、PPP2R5B和EPO。结论基于非参数正态图模型的差异网络分析方法在模拟实验中筛选差异基因的效果较好,在实例数据分析中筛选的肝细胞癌生存时间相关的枢纽基因都有较为明确的生物学功能。
Objective To evaluate the screening effect of differential network analysis based on non-paranormal graphical models, and to screen the hub genes affecting the survival time of patients with hepatocellular carcinoma using the differential network analysis model so as to predict the patients’ prognoses or serve as a new drug target. Methods The screening effect of differential network analysis using non-paranormal graphical models was evaluated by the area under the ROC curve(AUC) values in the simulation experiments. The gene expression data of hepatocellular carcinoma from 3 platforms and 5 datasets were downloaded from The Cancer Genome Atlas(TCGA) database, and then incorporated prior information such as pathways into differential network using non-paranormal graphical models to construct the differential network. The hub genes were chosen by the degree, the number of edges connected to the node. Results In the simulation experiments performed under four conditions, the AUC values of the non-paranormal graphical models ranged from 0.71 to 0.97. In the application of hepatocellular carcinoma data, 8 hub genes related to the survival time of hepatocellular carcinoma were screened respectively, which were IGF1, ErbB2, FGF9, GH2, cSH2, HSP90 AA1, PPP2 R5 B and EPO. Conclusions The simulation experiments of differential network analysis based on non-paranormal graphical models show good results. The biological function of hub genes affecting the survival time of patients with hepatocellular carcinoma is clearly defined.
作者
李晶
张奇
任雨冬
刘艳
LI Jing;ZHANG Qi;REN Yu-dong;LIU Yan(Department of Health Statistics,Harbin Medical University,Harbin,Heilongjiang 150081,China)
出处
《实用预防医学》
CAS
2019年第4期404-408,共5页
Practical Preventive Medicine
基金
国家自然科学基金(81172741)
关键词
差异网络
肝细胞癌
生存时间
枢纽基因
非参数正态图模型
differential network
hepatocellular carcinoma
survival time
hub gene
non-paranormal graphical model
