基于人工神经网络的当归-川芎药对整体功效与化学物质关联性分析  

Analysis on correlation between general efficacy and chemical constituents of Danggui-Chuanxiong herb pair based on artificial neural network

作　　者：李伟霞[1] 唐于平[1] 尚尔鑫[1] 郭建明[1] 黄美艳[1] 钱大玮[1] 段金廒[1] 

机构地区：[1]南京中医药大学江苏省方剂高技术研究重点实验室,江苏南京210046

出　　处：《中国中药杂志》2012年第19期2935-2942,共8页China Journal of Chinese Materia Medica

基　　金：国家科技支撑计划项目(2008BAI51B01);国家自然科学基金项目(30873235,81274058);江苏省研究生培养创新工程(CXZZ12_0627);2011年度高等学校博士学科点专项科研基金(20113237110010);江苏省高校自然科学重大基础研究项目(06KJA36022,07KJA36024,10KJA36023,12KJA36002);国家教育部“新世纪优秀人才支持计划”项目(NCET-09-0163);江苏高校优势学科建设工程项目(ysxk-2010)

摘　　要：目的:对当归-川芎(简称归芎)药对不同配比组方养血补血、活血化瘀和调经止痛效应进行综合评价,揭示其效应与物质的关联性及其程度。方法:首先将各效应指标进行标化,采用多指标综合指数法进行效应整合,将专家评分法与近10年来相关文献中对各效应指标的检测频次统计相结合,综合给出各指标的权重系数,然后将各指标的标化值乘以权重系数后相加即得总效应值;最后运用人工神经网络将归芎药对中7种主要效应成分和总芳香酸、总苯酞内酯的含量与其总效应分别进行拟合分析。结果:归芎药对1.5∶1水提物的总补血效应最好,1∶1醇提物的总活血化瘀效应最好,1.5∶1醇提物的总调经止痛效应最好;提取方法相同的情况下,当归提取物的总补血效应在一定程度上优于川芎提取物,川芎提取物的总活血化瘀作用和调经止痛作用在一定程度上优于当归提取物;配比相同的情况下,归芎药对水提物的总补血效应最好,醇提物的总活血化瘀作用和调经止痛效应最好。养血补血的效应物质主要为芳香酸,尤其是绿原酸和咖啡酸;芳香酸和苯酞内酯都是活血效应物质,尤其是绿原酸、洋川芎内酯I和H对活血效应影响较大,调经止痛的效应物质也为芳香酸和苯酞内酯,尤其是藁本内酯、咖啡酸、阿魏酸和洋川芎内酯I对调经止痛效应的影响较大。结论:归芎药对中芳香酸和苯酞内酯都对养血补血、活血化瘀和调经止痛功效具有贡献,芳香酸更偏重于养血活血,苯酞内酯更偏重于调经止痛。该研究通过多指标综合指数法对方药功效各测定指标进行整合,采用人工神经网络将方药成分含量与其整合效应分别进行拟合性相关分析,从而阐释方药不同功效的效应物质及其对整体功效的贡献大小,为方药复杂效应物质基础研究提供了新的思路与方法。Objective： To evaluate the effect of Danggui-Chuanxiong herb pair with different prescription proportions and formulas on nourishing and tonifying blood （NTB）, activating blood circulation and dissolving blood stasis （ADBS）, regulating menstruation and relieving pain （RMRP） in an all-round way, in order to reveal the correlation between the effect and constituents and their degrees. Method： First, the indexes of effect were standardized. All effects were integrated by using multi-index aggregative index method. The weight coefficients of each index were worked out by the combining the statistics of testing frequency of each effect index given by the specialist-scored method and in relevant literatures released in the latest 10 years. Then, the total effect values were obtained by multiplying the standardized value of each index with the weight coefficients and then adding their results together. Finally, a fitting analysis was made on the contents of seven major active components in Danggui-Chuanxiong herb pair, total aromatic acids and total phthalide lactones and their total effect by using the artificial neural network, respectively. Result： Water extracts of Danggui-Chuanxiong （1.5：1） had the best effect on total NTB; alcohol extracts of Danggui-Chuanxiong herb pair （1：1） had the best effect on total ADBS; alcohol extraction of Danggui-Chuanxiong herb pair （1.5：1） had the best effect on RMRP. By using the same extraction method, extracts of Danggui had better effect on total NTB than Chuanxiong; while extracts of Chuanxiong had better effect on total ADBS and RMRP than Danggui. With the same prescription proportion, water extracts of Danggui-Chuanxiong herb pair had the best effect on total NTB; alcohol extracts of Danggui-Chuanxiong herb pair had the best effect on total ADBS and RMRP. Aromatic acids were the main effective components for NTB, especially chlorogenic acid and caffeic acid. Both aromatic acids and phthalide lactones were active components for ADBS, particula

关 键 词：当归 川芎 药对 多指标综合指数法 人工神经网络 整体功效 化学物质 关联性分析 

分 类 号：R285[医药卫生—中药学]