论文部分内容阅读
以金银花药材为研究对象,收集40批有代表性的金银花药材,分别测定其色谱指纹图谱和抗流感病毒的IC50,再基于The Unscrambler软件,运用偏最小二乘法对色谱指纹图谱数据和IC50进行谱-效相关性分析,建立金银花药材抗流感病毒谱-效相关性评价模型,另取10批金银花药材对该模型进行适应性验证,以探讨基于偏最小二乘法建立金银花抗流感病毒谱-效相关性模型的适应性。研究结果得到,所建模型的校正集及交叉验证集的相关系数(R~2)分别为0.969 489,0.959 042,校正集及交叉验证集的均方根误差(RMSEC,RMSECV)分别为0.070 691,0.084 005;10批金银花药材的验证结果表明模型预测值和实测值的相对误差均在10%以内,其中80%相对误差在5%以内。结果表明,运用所建立的谱-效相关性模型可较好地通过金银花的指纹图谱数据评价其抗甲型流感病毒的生物活性。
Taking honeysuckle as the research object, 40 batches of typical honeysuckle medicinal materials were collected and their IC50s were determined respectively. Based on The Unscrambler software, partial least square method was used to analyze the chromatographic fingerprint data and IC50 -effective correlation analysis to establish the Honeysuckle herbs anti-influenza virus spectrum-efficacy correlation model, another 10 batches of honeysuckle herbs adaptive validation of the model to explore based on partial least square method to establish the honeysuckle anti-influenza virus spectrum-related Adaptability of sexual models. The results show that the correlation coefficients (R ~ 2) of the calibration set and cross validation set are 0.969489,0.959 042 respectively, and the root mean square error (RMSEC, RMSECV) of the calibration set and cross validation set are 0.070 691 , 0.084 005. The validation results of 10 batches of honeysuckle showed that the relative errors of the model predictive value and the measured value were within 10%, of which 80% relative error was within 5%. The results showed that the bioactivity of anti-influenza A virus could be evaluated by using the fingerprint-effective correlation model.