摘要
中药活性成分较为复杂,加之配伍后的药理活性多样,使得阐明化学成分间的活性成分及其协同作用较为困难,即使单味中药的成分间配伍关系研究也是一个庞大的工程。谱效关系是一种新兴的方法,目前已逐渐应用于单味药及复方药效物质基础、组分配伍、炮制机制、药效预测及工艺优化等方面的研究中。它是一种采用数据分析技术,将多维多息的指纹图谱数据与药理指标进行关联和分析的研究方法,是药学与数学的交叉领域。该方法的研究依赖于数据分析技术。客观合理的选择数据分析技术是必不可少的环节。该文对近年来数据分析技术在中药谱效关系中的应用进展进行了归类综述,其中包括,预测各成分与药效间关联度的分析方法:灰关联度分析法,人工神经网络,双变量相关分析法;阐明各成分对药效贡献率的分析方法:多元线性回归法,偏最小二乘回归分析法;简化数据结构寻找主要活性成分的分析方法:主成分分析法,典型相关分析法。并总结了目前常用的数据分析技术的特点,研究思路及其适用研究对象等。最终针对各分析技术存在的局限性进行了讨论,为谱效关系中数据分析技术的选择提出了建议。推荐合理地联用多种数据分析技术来建立较为理想的谱效关系模型,进而保证结果的客观性与真实性。
Because traditional Chinese medicine（ TCM） feature complicated active ingredients and diverse compatible herbs,it is difficult to define the active ingredients and their synergistic effect and even conduct a study on the compatibility between single TCM ingredients. The spectrum-effect relationship,as an emerging method,has been gradually applied in studies on the effect material foundation of compounds, component compatibility,processing mechanism,efficiency prediction and process optimization. The study method is a databased analytical techniques to associate the multidimensional fingerprint data with the pharmacological indexes as well as a crossing field between pharmacy and mathematics. The study relies on the data analysis technique and focuses on selecting objective and reasonable data analysis technique. In this paper,the authors summarized the advance in applications of data analysis technique in the spectrum-effect relationship,including,forecasting the analysis method on the correlation between ingredients and efficacy： grey relational analysis,artificial neural networks,bivariate correlations analysis. Defining the analysis methods on the contribution of each component to the efficacy： multiple linear regression,partial least squares; simplifying the analysis method for finding major active ingredients： principal component analysis,canonical correlation analysis. Besides,the characteristics of currently common data analysis technique,research ideas and their applicable research objects were summarized. At last,the authors discussed the limitations in various data analysis technique,gave suggestions for selecting data analysis technique in the spectrum-effect relationship and recommended the reasonable combination of a variety of data analysis technique,so as to establish an ideal spectrum-effect relationship model and ensure the objectivity and authenticity of the results.
出处
《中国实验方剂学杂志》
CAS
CSCD
北大核心
2015年第15期226-230,共5页
Chinese Journal of Experimental Traditional Medical Formulae
基金
国家自然科学基金面上项目(2013ZX09102019)
国家"重大新药创制"科技重大专项(2013ZX09102019)
黑龙江中医药大学创新人才基金项目(2012)
关键词
谱效关系
数据分析技术
综述
spectrum-effect relationship
data analysis technique
summary
