基于生物信息学研究氧化应激和铁死亡在脓毒症中的诊断价值  

Diagnostic value of oxidative stress and ferroptosis in sepsis based on bioinformatics analysis

作　　者：周金泉 李依慧 袁亚敏 马莉[2] ZHOU Jinquan;LI Yihui;YUAN Yaming;MA Li(The Second Clinical Medical School,Lanzhou University,Gansu Province,Lanzhou730030,China;Department of Critical Care Medicine,the Second Hospital,Lanzhou University,Gansu Province,Lanzhou730030,China)

机构地区：[1]兰州大学第二临床医学院,甘肃兰州730030 [2]兰州大学第二医院重症医学科,甘肃兰州730030

出　　处：《中国医药导报》2024年第22期27-33,共7页China Medical Herald

基　　金：甘肃省自然科学基金项目(22JR5RA965)。

摘　　要：目的探究氧化应激和铁死亡相关基因在脓毒症中的诊断价值。方法使用R语言筛选脓毒症数据集差异表达基因(DEG),并与氧化应激及铁死亡基因取交集确定候选基因并进行基因本体/京都基因和基因组数据库(GO/KEGG)通路分析。使用LASSO回归和SVM-RFE机器学习确定诊断基因建立诊断模型,纳入兰州大学第二医院2024年2月收治的3例新发脓毒症患者(脓毒症组)和3名健康者(对照组)的血液样本做进一步验证。结果从876个DEG中共获得10个候选基因,557个GO条目和32个KEGG通路被富集;机器学习后筛选出TXN、TIMP1并建立诊断模型,模型在训练集及验证集中曲线下面积均>0.9。脓毒症组血液样本TXN和TIMP1 mRNA表达量高于对照组(P<0.05)。结论TXN、TIMP1两基因模型能够较好地早期诊断脓毒症,可作为新的脓毒症生物标志物。Objective To explore the diagnostic value of genes related to oxidative stress and ferroptosis in sepsis.Methods The differentially expressed genes(DEG)of the sepsis training set were filtered using R language and intersected with genes related to oxidative stress and iron death for gene ontology/Kyoto Encyclogedia of Genes and Genomes(GO/KEGG)pathway analysis.Diagnostic genes were identified through LASSO regression and SVM-RFE machine learning to develop the diagnostic model.Validation involved analyzing blood samples from three recently diagnosed sepsis patients(sepsis group)admitted to the Second Hospital of Lanzhou University in February 2024,along with blood samples from three healthy individuals(control group).Results Following the analysis of 876 DEG,a set of ten candidate genes were identified,along with significant enrichment in 557 GO entries and 32 KEGG pathway.After machine learning,TXN and TIMP1 were screened out and a diagnostic model was established,the area value of the model under the curve of the training set and verification set were both>0.9.The expression levels of TXN and TIMP1 mRNA in blood samples of sepsis group were higher than those in control group(P<0.05).After machine learning,TXN and TIMP1 were screened out and a diagnostic model was established,the area value of the model under the curve of the training set and verification set were both>0.9.The expression levels of TXN and TIMP1 mRNA in blood samples of sepsis group were higher than those in control group(P<0.05).Conclusion The TXN and TIMP1 gene model can effectively diagnose sepsis in its early stages,making them potential new biomarkers for sepsis.

关 键 词：生物信息学 氧化应激 铁死亡 生物标志物 诊断模型 

分 类 号：R459.7[医药卫生—急诊医学]