基于知识学习挖掘乳腺癌与甲状腺癌的共享功能模块和核心基因

Identifying Breast Cancer and Thyroid Cancer Shared Functional Modules and Core Genes Based on Knowledge Learning Methods

目的基于基因多效性理论,利用转录组数据经生物分子网络分析方法探讨乳腺癌(BC)与甲状腺癌(TC)的共病分子机制。方法从肿瘤基因组图谱(TCGA)获取BC和TC的RNA-seq数据和临床信息,筛选两种疾病中癌症组与正常组的差异表达基因(DEGs),取两种疾病共同上下调交集的DEGs作为共享风险基因(CRGs)。以此作为种子,利用蛋白质-蛋白质互作知识扩充(PPI)构建共同风险基因网络。然后,应用Newman网络分解算法对所构建的基因网络划分功能模块和识别核心基因。最后,对功能模块及核心基因进行KEGG通路富集分析和生存分析以探讨这些功能模块和核心基因的生物学意义。结果共筛选出708个上调、428个下调CRGs并构建出一个包含2261个节点的共享风险基因网络。Newman算法提取了17个关键功能模块和105个核心基因。功能富集分析显示,功能模块主要与雌激素信号通路、Hippo信号通路等通路相关(P<0.05)。生存分析确定M3模块及其中的CD79A、GHR、RORB等基因与疾病预后相关。结论本研究基于基因多效性构建两种癌症的共享风险基因网络,识别了17个关键功能模块和105个核心基因,为揭示两种疾病的发病机制、发掘有效生物标志物提供参考。

Objective Based on the theory of gene pleiotropy,the molecular mechanism of breast cancer(BC)and thyroid cancer(TC)comorbidity was explored by using transcriptome data and biomolecular network analysis.Methods The RNA-seq and clinical data of BC and TC were acquired from the Cancer Genome Atlas(TCGA).The differentially expressed genes(DEGs)between the disease group and the normal group were obtained,which of the two diseases in common expression direction were selected as the common risk genes(CRGs).Using these as the seed,the common risk gene network was constructed by using the knowledge of protein-protein interaction(PPI)in HRPD database.Then,the Newman network decomposition algorithm was used to divide the gene network into significant functional modules and identify core genes.Finally,the pathway enrichment analysis based on Kyoto encyclopedia of gene and genome and survival analysis were carried out to explore the biological functions.Results A total of 708 up-regulated and 428 down-regulated CRGs were screened and a shared risk gene network with 2261 nodes was constructed.17 significant functional modules and 105 core genes were extracted by Newman algorithm.The results of functional enrichment showed that these significant modules were mainly related to complement and coagulation cascade,estrogen signal pathway,Hippo signal pathway and so on(P<0.05).Survival analysis confirmed that M3 modules(including CD79A,GHR,RORB and et al)were most related to the prognosis of diseases.Conclusion In this study,a shared risk gene network of two kinds of cancer was constructed based on gene polymorphism,and 17 significant functional modules and 105 core genes were identified,which play important roles for revealing the pathogenesis of the two diseases and exploring effective biomarkers.