应用PCA-ANN技术研究中药厚朴色谱指纹谱与抗氧化活性的关系  

The Relationship Between Chromatographic Fingerprints of Magnoliae officinalis and its Antioxidant Activity by PCA-ANN

作　　者：冯慧萍[1] 杨中林[2] 尹小英[1] 姚卫峰[1] 胡育筑[1] 

机构地区：[1]中国药科大学分析化学教研室,南京210038 [2]中国药科大学中医药教研室,南京210038

出　　处：《中国天然药物》2005年第6期377-381,共5页

基　　金：国家科技部与比利时弗拉芒大区科技合作项目"植物药指纹色谱质量控制的方法学和指导原则(BWS(BIL)02/06)"~~

摘　　要：目的:探索中药厚朴色谱指纹谱与抗氧化活性关系的研究方法。方法:建立浙江丽水产的15个厚朴样品的HPLC-DAD指纹图谱,对其共有峰的相对峰面积进行主成分分析,将前四个主成分的得分作为人工神经网络的输入;另采用化学发光法测定上述样品清除超氧阴离子和过氧化氢两种活性氧的能力,以清除率为50%时所需药物的浓度(IC50)作为神经网络的输出,构建厚朴色谱指纹图谱与其清除活性氧能力间关系的网络。用缺一交叉验证法检验网络的稳定性与预测精度。结果:清除超氧阴离子训练集的平均拟合偏差ASEr(拟合值与实验值之间的相对偏差)≤0.150,预测集的预报偏差PEr≤0.181;清除过氧化氢的ASEr≤0.110,PEr≤0.191。结论:用PCA-ANN技术构建的网络稳定性好,可准确预测厚朴清除活性氧能力的IC50值;此法可作为研究中药色谱指纹谱与抗氧化活性关系的一种方法。AIM： To explore the methodology for the relationship between chromatographic fingerprints of Magnoliae officinalis Rehd. et Wils. and its antioxidant activity. METHOD： The chromatographic fingerprints were established based on the HPLC-DAD analysis of 15 Magrtoliae offwinalis samples collected in Lishui, Zhejiang Province. The relative areas of characteristic peaks were disposed by principal components analysis; the scores of the first-four principle components were used as the input of artificial neural network. The scavenging reactive oxygen species ability of supemxide and hydrogen peroxide was evaluated by chemil technique. The data of IC50（the concentration of extracts causes 50% scavenging rate） were used as output of artificial neural network. The relationship was established by artificial neural network between chromatographic fingerprints of Magnoliae officinalis and its scavenging reactive oxygen ability. The stability and accuracy of predicted outputs were tested by the leave-one-out cross validation. RESULT： In supemxide analysis, the average simulated deviations of training samples ASEr （between the simulated data and the experimented values）≤ 0.150, the predicted deviation of predicting samples PEr≤ 0.181. In hydrogen peroxide analysis, ASEr≤0. 110,PEr≤0. 191. CONCLUSION： The network was stably established by PCA-ANN technique. It is possible to accurately predicted the IC50 values by ANN. This technique could be used as a method studying the relationship between chromatographic fingerprints and its antioxidant activity.

关 键 词：厚朴 色谱指纹图谱 清除自由基能力 主成分分析 神经网络 

分 类 号：R284.1[医药卫生—中药学] R965[医药卫生—中医学]