基于聚类、主成分和判别分析的海龙红外指纹图谱研究  

Infrared Fingerprint Analysis of Syngnathus' s Ethanol Extracts Coupled with Cluster Analysis,Principal Component Analysis and Discriminant Analysis

作　　者：吴伟建 王燕 王斌 任西杰 徐焕志 郑斌 杨会成 

机构地区：[1]浙江海洋学院食品与药学学院,浙江省海洋生物医用制品重点工程技术研究中心,浙江舟山316004 [2]浙江省海洋开发研究院,浙江舟山316021

出　　处：《中国药学杂志》2013年第18期1540-1545,共6页Chinese Pharmaceutical Journal

基　　金：国家自然科学基金资助项目(81001393);舟山市科技计划资助项目(2011C22060);海洋渔业科学与技术浙江省重中之重学科开放课题(20110223)

摘　　要：目的利用红外光谱指纹图谱结合数理统计分析方法建立一种海龙种类快速鉴别和质量控制方法。方法应用红外光谱技术分析33个不同来源的海龙样品乙醇提取物，通过聚类分析和相似度分析建立海龙乙醇提取物的红外光谱指纹图谱；利用主成分分析及判别分析进行海龙种类鉴别和质量控制研究。结果各海龙样品醇提物的红外光谱与所建立的指纹图谱共有模式比较相似度良好，符合中药指纹图谱要求；3种正品海龙及1种伪品海龙聚集为不同的类别，实现了不同种类海龙的快速鉴别；在主成分分析的基础上进行判别分析时，红外吸收峰3010．4～3006．5cm-1、1745．3～1741．4cm-1、1467．6—1463．7cm-1、1417．4～1413．6cm-1、1168．7—1164．8cm-1。、723．2～719．3cm-1在主成分1上的栽荷最大；吸收峰2925．5～2919．7cm-1、2854．2—2850．3Ocm-1在主成分2上的贡献最大，据主成分1和2建立的判别函数可用于区分3种正品海龙及其伪品，其正确率达到98．75％。结论本实验建立的乙醇提取物红外光谱指纹图谱结合主成分分析及判别分析可快速、准确地区分不同种类的海龙，可用于海龙的种类鉴别和质量控制。OBJECTIVE To establish a validated infrared spectrum method coupled with cluster analysis, principal component a- nalysis and diseriminant analysis for identifying different kinds of Syngnathus. METHODS Fourier transform infrared （FTIR） spec- trometer was performed on the ethanol extracts from 33 samples of Syngnathus collected from different geographical locations. Similarity analysis and cluster analysis were used to establish the reference fingerprint spectra, and principal component analysis and discriminant analysis were used for identifying different kinds of Syngnathus. RESULTS The experimental results indicated that the samples of same species had the similar characteristic peaks of infrared spectra and could be clustered together. Based on the result of principal component analysis, the absorption peaks（3010.4～3006.5cm-1、1745.3～1741.4cm-1、1467.6-1463.7cm-1、1417.4～1413.6cm-1、1168.7-1164.8cm-1、723.2～719.3cm-1） were the main characteristic peaks and were used to develop the discrimination function for distinguishing different species of Syngnathus. The accura- cy of the discrimination function could reach 98.75%. The established method in the experiment had a good stability, precision and re- producibility, and was a very reliable and useful method for distinguishing different species and assessment of the quality of Syn- gnathus. CONCLUSION The established IR fingerprints of Syngnathus＇s ethanol extracts coupled with ruathematical statistics theory could be applied to identify different kinds of Syngnathus. But more importantly, the research job provides a new thought for traditional Chinese medicine identification by using mathematical model.

关 键 词：海龙 醇提物 红外光谱 指纹图谱 聚类分析 主成分分析 判别分析 

分 类 号：R282[医药卫生—中药学]