基于全谱数据库计算指纹图谱相似度新方法的探索研究  

Exploration of a New Approach for Calculating Similarity of Fingerprints Based on the Whole Chromatogram

作　　者：刘云飞[1] 周利[1] 张春凤[1] 杨中林[1] 

机构地区：[1]中国药科大学天然药物活性物质与功能国家重点实验室,江苏南京210009

出　　处：《中医药学报》2012年第2期49-51,共3页Acta Chinese Medicine and Pharmacology

基　　金：国家自然科学基金重点项目(No.30730113);江苏高校优势学科建设工程资助项目

摘　　要：目的:建立一种以全谱数据库计算指纹图谱相似度的新方法。方法:从安捷伦工作站中导出全谱数据点;采用相关优化变形校正漂移的保留时间数据库;将校正后的数据库录入Excel中;采用"夹角余弦法"计算并比较10批药材校正前后的相似度。结果:以全谱数据库为研究对象,校正漂移的色谱峰保留时间后,计算所得药材的相似度普遍高于校正前的相似度;本文所建的"以全谱数据库计算中药指纹图谱相似度的新方法"与"以共有峰模式计算中药指纹图谱相似度的方法"相比,计算所得到的相似度值偏低。结论:以全谱数据库代替共有峰模式计算中药指纹图谱的相似度,避免了非共有峰信息的丢失,更能真实反映各药材指纹图谱间的相似性,计算结果客观,科学,合理。本文仅对该计算方法进行了初步探索。Objective： To establish a new approach to calculate the similarities of fingerprints based on the whole chromatogram.Methods： First chromatographic data points were exported from Chemstation.Then retention time shifts were corrected by correlation optimized warping.At last the unwarped and warped chromatographic data points were put in a sheet of Excel and similarities of ten samples were calculated with angle cosine method by using functions of Excel.Results： Similarities were widely improved after warping.But values were lower than those calculated by common peaks.Conclusion： It is more reasonable and authentic to calculate similarity with the whole chromatogram instead of common peaks,which can efficiently avoid the loss of important information.

关 键 词：全谱 共有峰 相似度 续断 质量控制 

分 类 号：R284.1[医药卫生—中药学]