基于熵值的尿路感染疾病基因网络的模块划分与生物学机制分析

Module Partition and Biological Mechanism Analysis of Genetic Network of Urinary Tract Infection Based on Entropy

尿路感染的病因比较复杂、且治愈后容易复发,因此,治疗尿路感染疾病已经在人们日常生活中成为了一大难题。本研究通过从Pubmed数据库获取与尿路感染疾病相关的基因共59个,然后运用String数据库建立尿路感染疾病的基因网络（有54个节点和337条边）,进而利用MCODE cluster、MCL cluster和Community cluster （glay） 3种常见的模块划分方法对尿路感染疾病基因网络进行模块识别,通过网络结构熵值进行评价,发现MCODE cluster方法划分出4个模块,网络熵值为3.179 4,在几种方法中熵值最小,最适于尿路感染疾病基因网络模块的识别。运用DAVID对尿路感染疾病的基因网络和MCODE cluster划分出的模块进行功能富集分析,分别得到了38和32条KEGG信号通路,且有70.7%的覆盖率,说明了MCODE的模块识别方法可以识别出疾病基因网络中与疾病生物学功能关系较为密切的基因,并可以找出可以代表该基因网络大部分生物学功能的模块网络。这为后期研究“药物-靶点-疾病”间关联提供了一种可行方案。

Urinary tract infection has complex causes and is easy to relapse, so the treatment of urinary tract infection has become a major problem in daily life. In this study, 59 genes related to urinary tract infection were obtained from the Pubmed database, and the String database was used to establish the gene network of urinary tract infection. Then, MCODE cluster, MCL cluster and Community cluster were used to identify the genetic network of urinary tract infection. The network entropy was evaluated by the network structure. We found that the MCODE cluster method divided four modules with a network entropy of 3.179 4, which was the minimum entropy in three method and the most suitable for urinary tract infection disease gene network module identification. DAVID was used to characterize the gene network of the urinary tract infection and the modules enriched by the MCODE cluster. 38 and 32 KEGG pathways were obtained, respectively, and 70.7% coverage was obtained, which indicated that the module recognition method of MCODE could identify tile genes in the disease gene network that were more closely related to the biological function of the disease and could identify the module network that can represent most of the biological functions of the gene network. This research a viable option for the later study of the association between ＂drug-target-disease ＂.