关联分析和互信息熵联合探究《中国药典》胃病相关中药成方制剂用药规律  

Exploration on prescription regularity of Chinese patent medicines against stomach diseases in Chinese Pharmacopoeia based on correlation analysis and mutual information entropy

作　　者：高苏雅 李怡 姚鉴玲 周芷熠 钟祥健 马嘉慕 李仙仙 孙梦豫 齐晓丹 折改梅[1] GAO Su-ya;LI Yi;YAO Jian-ling;ZHOU Zhi-yi;ZHONG Xiang-jian;MA Jia-mu;LI Xian-xian;SUN Meng-yu;QI Xiao-dan;SHE Gai-mei(School of Chinese Materia Medica,Beijing University of Chinese Medicine,Beijing 100029,China;National Engineering Technology Research Center for Gelatin-based Traditional Chinese Medicine,Shandong Key Laboratory of Gelatine TCM Research and Development,Dong’e Ejiao Co.,Ltd.,Liaocheng 252200,China)

机构地区：[1]北京中医药大学中药学院,北京100029 [2]东阿阿胶股份有限公司山东省胶类中药研究与开发重点实验室国家胶类中药工程技术研究中心,山东聊城252200

出　　处：《中草药》2023年第18期6006-6016,共11页Chinese Traditional and Herbal Drugs

基　　金：泰山产业领军人才工程资助(tscx202211148)。

摘　　要：目的基于关联分析和互信息熵对《中国药典》2020年版一部与胃病相关的中药成方制剂进行药物配伍应用规律研究。方法锁定“胃”“脘”“宽中”“纳呆”“呕恶”“食积”等为检索词,收集《中国药典》中胃病相关成方制剂,结合药智数据网、蒲标网及中国知网(CNKI)辅助检索其药材相关信息,使用Microsoft Excel 2019软件建立数据库。运用SPSS Modeler 18.0的Apriori算法对高频中药进行关联规则分析,并使用Cytoscape 3.9.1完成网络可视化。用互信息熵(mutual information entropy,MIE)量化药对关联程度,选择高MIE药对在Cytoscape 3.9.1中构建药材配伍网络,并采用Glay算法对网络节点进行聚类分析。并将以上2种方法进行组合分析,先用Apriori算法收集所有核心中药组合,再计算并筛选出高MIE值药对。结果《中国药典》2020年版一部共收录263个胃病相关成方制剂,剂型以丸剂、胶囊剂、片剂为主;涉及352味药材,主要包括清热药、补虚药、祛风湿药等,药味以苦、辛、甘为主,药性以温、寒居多。其中有13味名贵中药,34味毒性中药。在所有药材中,使用频率较高的中药有甘草、陈皮、木香、茯苓、白术。以使用频率≥4%的中药进行关联分析并筛选得到40条强关联规则。MIE分析得出联系较强的药对有六神曲-麦芽、山楂-麦芽、山楂-六神曲等,聚类分析得到12个网络集群,各网络集群包含不同中药成方制剂。Apriori-MIE组合分析得到新的重要关联规则厚朴+茯苓→广藿香。结论山楂-六神曲-麦芽是临床治疗胃病的重要潜在药对,通过Apriori关联分析、MIE分析及其组合关联分析,全面深入挖掘了胃病相关中药成方制剂的配伍规律,为胃病临床治疗及相关中药成方制剂的开发提供参考。Objective To investigate the drug compatibility rules of Chinese patent medicines(CPMs)against stomach diseases in volume 1 of Chinese Pharmacopoeia(2020 edition)based on correlation analysis and mutual information entropy.Methods Focusing on“stomach”“epigastrium”“eliminating flatulence”“anorexia”“nausea and vomiting”,and“dyspepsia”as keywords in Chinese Pharmacopoeia were collected to find CPMs against stomach diseases and the relevant information of the Chinese medicinal materials was retrieved by combining Yaozhidata.com,Pubiao website and China National Knowledge Infrastructure(CNKI),and the database was established using Microsoft Excel 2019 software.The Apriori algorithm of SPSS Modeler 18.0 was used to analyze the association rules of high-frequency traditional Chinese medicine(TCM),and Cytoscape 3.9.1 was used to complete the network visualization.Mutual information entropy(MIE)was used to quantify the correlation degree of drug pairs.Drug pairs with high MIE were selected to construct a drug compatibility network in Cytoscape 3.9.1,and Glay algorithm was used to perform cluster analysis on the network nodes.The above two methods were combined and analyzed.Apriori algorithm was used to collect all the core TCM combinations,and then high MIE value drug pairs were calculated and screened.Results A total of 263 CPMs were included in volume 1 of Chinese Pharmacopoeia(2020 edition),most of which were pills,capsules and tablets.A total of 352 kinds of medicinal materials were involved,mainly including heat-clearing herbs,tonifying deficiency herbs and dispelling wind-dampness herbs,etc.These herbs tended to be bitter,pungent and sweet in flavor but warm and cold in nature.A total of 13 rare TCMs and 34 poisonous TCMs were included.In all these herbs,Gancao(Glycyrrhizae Radix et Rhizoma),Chenpi(Citri Reticulatae Pericarpium),Muxiang(Aucklandiae Radix),Fuling(Poria)and Baizhu(Atractylodis Macrocephalae Rhizoma)were identified as key TCMs with high frequency.A total of 40 strong association

关 键 词：《中国药典》 胃病 药物配伍 关联规则 互信息熵 山楂 六神曲 麦芽 

分 类 号：R285[医药卫生—中药学]