基于Apriori和MIE算法分析关节炎中药制剂用药规律及核心药对的潜在作用机制  

Analysis of medication rules and potential mechanism of core drug pairs of traditional Chinese medicine preparation for arthritis based on Apriori and MIE algorithm

作　　者：翟桐 杜佳璐 任雪阳 孙梦豫 黎明霞 刘薇 和映玉 李仙仙 邓清月 折改梅[1] ZHAI Tong;DU Jialu;REN Xueyang;SUN Mengyu;LI Mingxia;LIU Wei;HE Yingyu;LI Xianxian;DENG Qingyue;SHE Gaimei(College of Traditional Chinese Medicine,Beijing University of Chinese Medicine,Beijing 100029,China)

机构地区：[1]北京中医药大学中药学院,北京100029

出　　处：《中草药》2024年第9期3027-3040,共14页Chinese Traditional and Herbal Drugs

基　　金：国家自然科学基金面上项目(82074121);国家自然科学基金青年科学基金项目(82104531)。

摘　　要：目的基于Apriori和MIE算法分析并总结《中国药典》2020年版中关节炎成方和单味制剂的用药规律。方法检索《中国药典》2020年版一部中治疗关节炎的成方制剂和单味制剂的信息。运用WPS Office Excel软件建立数据库,应用IBM SPSS Modeler 18.0软件中Apriori算法、MIE算法和Cytoscape 3.9.1进行关联规则分析、药对相关性分析等。选取MIE分析和关联规则分析中置信度和支持度相对较高的核心药物,进行基因本体(gene ontology,GO)功能富集和京都基因与基因组百科全书(Kyoto encyclopedia of genes and genomes,KEGG)通路富集分析,并进行成分-靶点-通路分析。结果共收集102个成方制剂和8个单味制剂,涉及中药271味,包含毒性中药42味,依频次高低排序前6的中药为当归、红花、乳香、没药、防风和独活。功效以祛风湿、补虚、活血化瘀为主,归经以肝经、脾经和肾经为主,四气以温为主,五味以辛为主。其中大毒中药6味,小毒中药19味,有毒中药17味,依频次高低排序前4的毒性中药为制川乌、制草乌、马钱子、樟脑。关联规则分析得到“没药-乳香”“乳香-没药-当归”等配伍最常见。MIE分析得到“党参-白芍”“狗脊-淫羊藿”等强相关潜在药对。其中,MIE分析和关联规则中支持度与置信度较高的药对为“当归-川芎”,药组为“当归-川芎-白芷”。网络药理学分析显示主要潜在活性成分为β-谷甾醇、豆甾醇、杨梅酮、阿魏酸等,核心靶点为PTGS2、HSP90AB1、CASP3、Bcl2、PTGS1、Bax等,主要涉及包含癌症相关通路、花生四烯酸代谢通路、亚油酸通路等。集群分析成方制剂以治疗风寒湿痹症为主。结论关节炎为正虚邪实证,治疗应以扶正驱邪、宣通脏腑、调达经络为主。毒性中药使用较多且用量较大,可增强疗效,但应该严格控制用量,防止中毒现象的发生。核心药对“当归-川芎”、核心药组“当归-川芎-白�Objective To analyze and summarize the medication rules of prescription and single preparations for the treatment of arthritis contained in the Chinese Pharmacopoeia(2020 edition)by Apriori and MIE algorithm.Methods To retrieve information on prescription and single flavor preparations for the treatment of arthritis in Part 1 of the Chinese Pharmacopoeia(2020 edition).WPS Office Excel was used to establish database.Apriori and MIE algorithm of the IBM SPSS Modeler 18.0 software and Cytoscape 3.9.1 were used to analyze the association rules,correlation of couplet medicines and,etc.MIE analysis and association rules were used to select core drugs with relatively high confidence and support,and gene ontology(GO)functional enrichment analysis,Kyoto encyclopedia of genes and genomes(KEGG)pathway enrichment analysis,and component-target-pathway analysis were performed.Results A total of 102 prescription preparations and eightsingle preparations were collected,including 271 traditonal Chinese medicines(TCMs),of which 42 were toxic TCMs.The top six TCMs ranked according to the frequency were Danggui(Angelicae Sinensis Radix),Honghua(Carthami Flos),Ruxiang(Olibanum),Moyao(Myrrha),Fangfeng(Saposhnikoviae Radix),and Duhuo(Angelicae Pubescentis Radix).The main effects are dispelling rheumatism,tonifying deficiency,promoting blood circulation and removing blood stasis.The main meridians are liver meridian,spleen meridian,and kidney meridian.The four qi are mainly warm,and the five flavors are mainly pungent.There are six large toxic TCMs,19 small toxic TCMs,and 17 toxic TCMs.The top four toxic TCMs ranked according to frequency were Zhichuanwu(Aconiti Radix Cocta),Zhicaowu(Aconiti Kusnezoffii Radix Cocta),Maqianzi(Strychni Semen)and camphor.Through association rule analysis,the compatibility of“Myrrha-Olibanum”and“Olibanum-Myrrha-Angelicae Sinensis Radix”are the most common things.Through MIE analysis,“Dangshen(Codonopsis Radix)-Baishao(Paeoniae Radix Alba)and Gouji(Cibotii Rhizoma)-Yinyanghuo(Epimedii Folium)”are

关 键 词：关节炎 数据挖掘 用药规律 毒性中药 MIE 网络药理学 当归-川芎 当归-川芎-白芷 

分 类 号：R285[医药卫生—中药学]