基于机器学习与分子动力学模拟挖掘天南星治疗重度抑郁症的潜在抗炎靶点和关键活性成分  

Exploration of potential anti-inflammatory targets and key ingredients of Arisaematis Rhizoma in treatment of major depressive disorder based on machine learning and molecular dynamics simulation

作　　者：郑远腾 夏漓 张秀军 靳梦 ZHENG Yuanteng;XIA Lijie;ZHANG Xiujun;JIN Meng(School of Psychology and Mental Health,North China University of Science and Technology,Tangshan 063200,China;School of Public Health,School of Psychology and Mental Health,North China University of Science and Technology,Tangshan 063210,China;Key Laboratory of Drug Screening Technology,Shandong Academy of Sciences,Biology Institute,Qilu University of Technology(Shandong Academy of Sciences),Jinan 250103,China)

机构地区：[1]华北理工大学心理与精神卫生学院,河北唐山063200 [2]华北理工大学心理与精神卫生学院、公共卫生学院,河北唐山063210 [3]齐鲁工业大学(山东省科学院)生物研究所,山东省科学院药物筛选技术重点实验室,山东济南250103

出　　处：《中草药》2024年第15期5174-5188,共15页Chinese Traditional and Herbal Drugs

基　　金：山东省重点研发计划(2022CXGC020515);泰山学者青年专家项目(tsqn202312233);山东省自然科学基金项目(ZR2021ZD29)。

摘　　要：目的利用生物信息学的方法结合多种机器学习模型,寻找天南星治疗重度抑郁的潜在抗炎靶点,并从免疫炎症角度探讨其发病机制。采用分子对接、药物相似性评估和分子动力学模拟,筛选天南星治疗重度抑郁的关键活性成分,以期对相关治疗药物与治疗策略的开发提供参考。方法利用CIBERSORT与基因集变异分析(gene set variation analysis,GSVA)验证免疫炎症假说,并寻找免疫标志物。基于TCMSP、GEO和Swiss Target Prediction数据库,获得天南星潜在作用靶点与重度抑郁相关靶点,并利用费舍尔精确检验探究相关靶点的关联性。采用加权基因共表达网络(weighted gene co-expression network analysis,WGCNA)分析获得与免疫炎症相关的基因共表达模块。通过将炎症相关基因、天南星潜在作用靶点、重度抑郁相关靶点、与免疫炎症相关的绿松石模块(Turquoise)取交集获得天南星治疗重度抑郁症的候选抗炎靶点。采用基因本体(gene ontology,GO)和京都基因与基因组百科全书(Kyoto encyclopedia of genes and genomes,KEGG)富集分析探究候选抗炎靶点涉及的生物学过程和相关通路。构建8个机器学习模型,以筛选潜在抗炎靶点。利用受试者工作特征(receiver operator characteristic,ROC)曲线与相关分析筛选出潜在抗炎靶点并进行外部验证。使用分子对接与药物相似性评估筛选天南星的关键活性成分,并采用分子动力学模拟验证分子对接与药物相似性评估的结果。结果正常组与重度抑郁患者的炎症得分存在显著性差异(P<0.05),CD8 T细胞为重度抑郁患者的免疫标志物。天南星潜在作用靶点和重度抑郁相关靶点具有关联性,与免疫炎症最相关的基因共表达模块为绿松石模块。候选抗炎靶点具有8个,涉及的生物学过程为炎症反应调节,相关通路为Fc-RI信号通路与T细胞受体信号通路。随机森林为最优模型,且纳入的靶点都具有较高�Objective To identify potential anti-inflammatory targets in treatment of major depressive disorder and explore its pathogenesis from the perspective of immune inflammation,bioinformatics methods combined with multiple machine learning models were utilized.Molecular docking,drug similarity assessment,and molecular dynamics simulation were employed to screen the key active ingredients of Tiannanxing(Arisaematis Rhizoma),in order to provide reference for the development of related therapeutic drugs and strategies.Methods CIBERSORT and gene set variation analysis(GSVA)were used to validate the immune inflammation hypothesis and identify immune biomarkers.Based on the TCMSP,GEO,and Swiss Target Prediction databases,potential targets of Arisaematis Rhizoma and targets related to major depressive disorder were obtained.Furthermore,the association between potential targets of Arisaematis Rhizoma and targets related to major depressive disorder was explored using Fisher’s exact test.Weighted gene co-expression network analysis(WGCNA)was employed to obtain gene co-expression modules most associated with immune inflammation.Anti-inflammatory candidate targets were obtained by intersecting inflammation-related genes,Arisaematis Rhizoma potential targets,major depressive disorder-related targets,and Turquoise module most associated with immune inflammation.Gene ontology(GO)and Kyoto encyclopedia of genes and genomes(KEGG)enrichment analysis were conducted to explore the biological processes and pathways involved in the anti-inflammatory candidate targets.Eight machine learning models were constructed to screen potential anti-inflammatory targets.Potential anti-inflammatory targets were selected using receiver operator characteristic(ROC)curves and correlation analysis.Additionally,external validation was conducted for the potential anti-inflammatory targets.Finally,molecular docking and drug similarity assessment were used to screen key active ingredients of Arisaematis Rhizoma.The molecular dynamics simulation was employe

关 键 词：天南星 重度抑郁 机器学习 分子对接 分子动力学模拟 8 11 14-二十二碳三烯酸甲酯 花生四烯酸5-脂氧合酶 

分 类 号：Q811.4[生物学—生物工程] R285[医药卫生—中药学] TP18[医药卫生—中医学]