基于计算机文本挖掘的头颈部肿瘤放化疗后口腔粘膜炎基因功能富集及药物治疗分析

Analysis of Gene Function Enrichment and Drug Treatment of Oral Mucositis after Radiotherapy and Chemotherapy for Head and Neck Tumor Based on Computer Text Mining

目的:利用计算机软件对已经公开的数据库进行挖掘和分析,确定口腔粘膜炎(Oral mucositis, OM)和伤口愈合相关的基因、蛋白、信号通路,对治疗OM可能有效的潜在药物进行预测和探索。方法:通过文本挖掘软件pubmed2ensembl确定与OM、伤口愈合相关的基因,去重后,用Venny 2.0筛选他们基因的交集,DAVID对交集基因进行基因功能的富集分析和京都基因与基因组百科全书(Kyoto Encyclopedia of Genes and Genomes, KEGG)通路分析。使用STRING对富集结果进行蛋白质–蛋白质相互作用的网络分析,Cycstcape进行蛋白质–蛋白质相互作用的模块分析。GEPIA进一步筛选模块分析的结果。最后,使用DGIdb进行分析,得到药物–基因相互作用的结果。结果:通过文本挖掘去重后,得到372个交集基因,对这些基因的生物学过程进行功能富集分析,并分析KEGG通路,通过蛋白质–蛋白质相互作用分析和模块分析,进一步筛选出37个相关基因。筛选出箱式图与生存率同时存在统计学差异的3个基因。最后,筛选出与之对应的38种药物。结论:使用文本挖掘、基因功能富集分析、分子途径等生物信息学工具,可以发现潜在的治疗OM的药物,为头颈部肿瘤放化疗后的放射性黏膜炎提供了新的线索。

Objective: Use computer software to mine and analyze the published databases to determine the genes, proteins and signal pathways related to Oral mucositis (OM) and wound healing, and to predict and explore potential drugs that may be effective in the treatment of OM. Method: Through the text mining software pubmed2ensembl, the genes related to OM and wound healing were identified. After de duplication, their gene intersections were screened with Venny 2.0. David enriched and analyzed the gene functions of the genes and the path analysis of Kyoto Encyclopedia of genes and genomes (KEGG). String was used to analyze the network of protein-protein interaction and cycstcap was used to analyze the module of protein-protein interaction. Gepia further screened the results of module analysis. Finally, dgidb was used to analyze the drug gene interaction. Result: After text mining and de duplication, 372 intersecting genes were obtained. The biological process of these genes was analyzed by function enrichment, and KEGG pathway was analyzed. Through protein interaction analysis and module analysis, 37 related genes were further screened. Three genes with statistically significant difference between box plot and survival rate were screened out. Finally, 38 corresponding drugs were screened out. Conclusion: Using bioinformatics tools, such as text mining, gene function enrichment analysis, molecular pathway and so on, we can find potential drugs for the treatment of OM.