基于数据挖掘和网络药理学技术的古代治疗消渴方剂的组方规律及其机制分析  

Analysis of the Grouping Pattern of Ancient Prescriptions for Xiaoke and Their Mechanisms Based on Data Mining and Network Pharmacology

作　　者：孙蕊 范颖[1] SUN Rui;FAN Ying(Liaoning University of Traditional Chinese Medicine,Shenyang 110847,Liaoning,China)

机构地区：[1]辽宁中医药大学,辽宁沈阳110847

出　　处：《辽宁中医药大学学报》2023年第2期73-78,共6页Journal of Liaoning University of Traditional Chinese Medicine

基　　金：国家自然科学基金面上项目(81273653)。

摘　　要：目的运用数据挖掘技术分析古代治疗消渴方剂的组方配伍规律,基于网络药理学方法分析核心组方药物的药理机制。方法以“消渴”为关键词,收集古代医案汇编书籍中的古代医案,检索《临证指南医案》《名医类案》等书中治疗消渴的方剂,建立方-药数据库。运用古今医案平台、IBM SPSS Modeler 18.0数据挖掘软件、IBM SPSS Statistics 25.0统计软件进行频次统计、关联规则分析、系统聚类分析和主成分分析寻求组方中的核心药物。借助中药系统药理学数据库收集核心药物的活性成分及其作用靶点;利用Cytoscape 3.7.3软件进行网络构建及可视化分析;将有效靶点导入STRING数据库,构建蛋白相互作用网络分析;利用DAVID数据库进行GO注释分析与KEGG通路分析。结果共筛选出254个方剂,涉及240味中药;使用频次在25以上的中药有27味,其中茯苓的使用频次最高;关联规则表明药对茯苓-泽泻,茯苓-山茱萸,支持度最高;聚类分析可将中药分为4类,其中第一类包括:生地黄、茯苓、泽泻、牡丹皮、山药、山茱萸、桂枝和附子;主成分分析显示第一主成分成员与聚类分析结果相同。综合频数、关联规则和聚类、主成分分析结果,在古代治疗消渴方剂中生(熟)地黄、茯苓、泽泻、牡丹皮、山药、山茱萸、桂枝和附子(金匮肾气丸)占据着核心地位。现代中药化学成分研究发现肾气丸治疗消渴的关键成分为槲皮素、山柰酚、豆甾醇、β型谷甾醇等,核心靶点为蛋白激酶Bα、白细胞介素-6、肿瘤抗原p53、肿瘤坏死因子。GO注释富集通路为蛋白结合,KEGG富集通路为癌症通路。结论通过数据挖掘,得出古方治疗消渴最常用方剂为金匮肾气丸及其加减化裁,其核心组成茯苓、泽泻、生地黄等药物,其核心成分为槲皮素、山柰酚等,这些成分通过蛋白结合通路作用于蛋白激酶Bα靶点、IL-6靶点等,为古方治疗�Objective To analyze the compatibility law of ancient prescription for treating Xiaoke by using data mining technology,and analyze the pharmacological mechanism of core prescription drugs based on network-pharmacology method.Methods Taking“Xiaoke”as the keyword,collecting the ancient medical records in the ancient medical records compilation books,searching the prescriptions for treating Xiaoke in the clinical evidence guide medical records,famous medical cases and other books,establishing a cubic-drug database.YIANKB,IBM SPSS modeler 18.0 and IBM SPSS statistics 25.0 will be used for frequency statistics with association rule analysis,system cluster analysis and principal component analysis to find the core drugs in the prescription.Collecting the active components and action targets of core drugs with the help of TCMSP;Using Cytoscape 3.7.3 for network construction and visual analysis.The effective targets will be imported into STRING to construct protein-protein interaction network analysis.Using DAVID for GO annotation analysis and KEGG path analysis.Results A total of 254 prescriptions were screened,involving 240 traditional Chinese medicines.27 traditional Chinese medicines were used more than 25 times,among which Fuling(Poria)showed the most frequently.Association rules showed that the drug had the highest support for Fuling(Poria)-Zexie(Alismatis Rhizoma)and Fuling(Poria)-Shanzhuyu(Cornifructus).Traditional Chinese medicine can be divided into four categories by cluster analysis,with the first category of Shengdihuang(Rehmanniae Radix),Fuling(Poria),Zexie(Alismatis Rhizoma),Mudanpi(Moutan Cortex),Shanyao(Dioscoreae Rhizoma),Shanzhuyu(Cornifructus),Guizhi(Cinnamomi Ramulus)and Fuzi(Aconm Lateralis Radix Praeparaia);which was as same as principal component analysis result.Based on the results of comprehensive frequency,association rules and clustering and principal component analysis,Shengdihuang(Rehmanniae Radix)/Shudihuang(Rehmanniae Radix Praeparata),Fuling(Poria),Zexie(Alismatis Rhizoma),Mudanpi(Mou

关 键 词：古代方剂 消渴 数据挖掘 网络药理学 组方规律 药理机制 

分 类 号：R285.5[医药卫生—中药学]