目的 通过生物信息学方法筛选并验证类风湿性关节炎(rheumatoid arthritis,RA)相关的铁死亡特征基因及预测作用中药.方法 从GEO数据库下载RA和健康对照者关节滑膜组织的基因芯片数据.通过FerrDb数据库获取RA相关的铁死亡差异表达基因(ferroptosis differential expression gene,FDEG),通过基因本体论(Gene Ontology,GO)和京都基因与基因组百科全书(Kyoto Encyclopedia of Genes and Genomes,KEGG)信号通路富集分析,探讨FDEG的生物学内涵和信号通路.再借助机器学习获取特征FDEG,结合外部验证集分析判断其诊断能力,并通过symMap数据库反向预测可以作用这些特征基因的中药.结果 从GEO数据库中下载得到铁死亡相关基因484个,从训练集基因列表中筛选出具有显著差异的125个FDEG.GO富集分析发现,这125个FDEG可以从多种生物学角度共同发病.KEGG富集分析揭示了它们在铁死亡、自噬、肿瘤以及缺氧诱导因子-1(hypoxia-inducible factor,HIF-1)、叉头框蛋白 O1(forkhead box protein O1,FoxO1)、白细胞介素-17(interleukin-17,IL-17)和肿瘤坏死因子(tumor necrosis factor,TNF)等信号通路中发挥作用.机器学习获得了 ATF3、AKR1C1、RRM2这3个RA铁死亡特征基因,进一步分析铁死亡特征基因在训练集和验证集中的表达水平,结果显示,AKR1C1、RRM2基因在RA铁死亡途径中具有一定的诊断效能.再对特征铁死亡基因进行基因集富集分析(gene set enrichment analysis,GSEA)和基因集变异分析(gene set variation analysis,GSVA),发现AKR1C1、ATF3和RRM2主要与代谢、合成等生物学过程密切相关.最终通过symMap数据库预测了可以作用这3个铁死亡特征基因的中药,发现化痰、理气类中药占比最大.结论 AKR1C1、RRM2基因在RA铁死亡途径中具有一定的诊断效能,同时理气、化痰类中药可以作为抗RA铁死亡途径的参考.
Bioinformatics-based Analysis of Ferroptosis Characteristic Genes in Rheumatoid Arthritis and Prediction of Traditional Chinese Medicine Intervention Potential
Objective To screen and verify the rheumatoid arthritis(RA)-related ferroptosis characteristic genes by bioinformatics methods,and predict the traditional Chinese medicine that could act on these feature genes.Methods Microarray data of joint synovial tissue of RA and healthy controls were downloaded from Gene Expression Omnibus(GEO)database,and the RA-related ferroptosis dif-ferential expression gene(FDEG)were obtained through the FerrDb database,and the biological connotations and signaling pathways of FDEG were explored through Gene Ontolog(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)analysis.Machine learning was used to obtain the feature FDEG,and the diagnostic ability was judged by combining it with the external validation set analysis.The symMap database was used to reverse predict the traditional Chinese medicine that could act on these feature genes.Results A total of 484genes related to ferroptosis were downloaded and 125 FDEG with significant differences were selected from the list of genes in the train-ing set.GO enrichment analysis revealed that these 125 FDEG could be co-morbid from multiple biological perspectives.KEGG analysis revealed their roles in ferroptosis,autophagy,tumors,and hypoxia-inducible factor(HIF-1),forkhead box protein O1(FoxO1),in-terleukin-17(IL-17),and tumor necrosis factor(TNF)signaling pathways.Machine learning obtained three RA ferroptosis character-istic genes,such as ATF3,AKR1C1 and RRM2.After analyzing the expression levels of the ferroptosis characteristic genes in the training and validation sets,it was found that AKR1C1 and RRM2had certain diagnostic efficacy in the RA Ferroptosis pathway.Then,gene set enrichment analysis(GSEA)and gene set variation analysis(GSVA)were performed on the ferroptosis characteristic genes,and it was found that AKR1C1,ATF3,and RRM2 were mainly closely related to metabolism,synthesis,and other biological processes.Finally,the symMap database predicted that the three ferroptosis characteristic genes could be affected by traditional Chinese medicines,and it was found that the phlegm-reducing and qi-regulating traditional Chinese medicine accounted for the largest proportion of the total.Conclu-sion AKR1C1 and RRM2genes have certain diagnostic efficacy in the RA ferroptosis pathway,and qi-regulating and phlegm-reduc-ing traditional Chinese medicine can be used as a reference for the anti-RA ferroptosis pathway.
Rheumatoid arthritisFerroptosisCharacterized genesTraditional Chinese medicine
万方数据
维普
