物理化学学报 >> 2015, Vol. 31 >> Issue (5): 893-898.doi: 10.3866/PKU.WHXB201503193

理论与计算化学 上一篇    下一篇

基于范数指数定量构效关系预测β-环糊精络合常数

钱海城1, KANWAL Shahid1, 贾青竹2, 王强1, 计惠芬1, 朱志臣3, 夏淑倩4, 马沛生4   

  1. 1 天津科技大学材料科学与化学工程学院, 天津300457;
    2 天津科技大学海洋科学与工程学院, 天津300457;
    3 天津城市建设学院理学院, 天津300384;
    4 天津大学化工学院, 天津300072
  • 收稿日期:2015-02-09 修回日期:2015-03-19 发布日期:2015-05-08
  • 通讯作者: 王强 E-mail:wang_q@tust.edu.cn
  • 基金资助:

    国家自然科学基金(21306137, U1162104)资助项目

Norm Index-Based Quantitative Structure-Activity Relationship to Predict β-Cyclodextrin Complex Binding Constants

QIAN Hai-Cheng1, KANWAL Shahid1, JIA Qing-Zhu2, WANG Qiang1, JI Hui-Fen1, ZHU Zhi-Chen3, XIA Shu-Qian4, MA Pei-Sheng4   

  1. 1 School of Material Science and Chemical Engineering, Tianjin University of Science and Technology, Tianjin 300457, P. R. China;
    2 School of Marine Science and Engineering, Tianjin University of Science and Technology, Tianjin 300457, P. R. China;
    3 School of Science, Tianjin University of Urban Construction, Tianjin 300384, P. R. China;
    4 School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, P. R. China
  • Received:2015-02-09 Revised:2015-03-19 Published:2015-05-08
  • Contact: WANG Qiang E-mail:wang_q@tust.edu.cn
  • Supported by:

    The project was supported by the National Natural Science Foundation of China (21306137, U1162104).

摘要:

环糊精在药剂学领域中是一类重要的包结化合物, 其中络合物稳定常数(logK)是一个关键评价参数. 本研究基于扩展距离矩阵提出了一组范数指数, 利用多种计算方法构建了系列定量构效关系模型, 并对233种化合物与β-环糊精的logK进行了计算预测. 计算结果表明基于扩展距离矩阵范数建立的系列定量构效关系模型均能较好预测logK; 其中利用最小二乘-支撑向量机方法建立的模型预测效果最好, 其预测结果的相关性系数R和留一、留十交叉验证相关性系数(QLOO,QLTO)分别为0.9587、0.8775和0.8732. 与文献方法对比结果表明, 本工作的预测结果在准确性和稳定性上有着显著的改善, 且能分辨同分异构体. 本课题组前期研究成果和本项工作表明基于范数指数构建的定量构效关系不仅适用于计算化合物的基础物理化学性质, 还能应用到化学反应过程相关常数的准确预测.

关键词: β-环糊精, 范数指数, 络合物稳定常数, 从头计算法, 构效关系

Abstract:

Cyclodextrins (CDs) are widely used in the pharmaceutical industry, and the complex stability constant (logK) is an important evaluation target for CD inclusion complexes. In this work, the structures of the inclusion complexes of 233 compounds with β-cyclodextrin (β-CD) were investigated by the quantitative structure-activity relationship (QSAR) method based on a new set of norm indexes proposed by our group. Here, using several arithmetic approaches, a set of QSAR models based on these new norm indexes were developed to predict the logK values of the β-CD complexes. The results showed that all of the norm indexbased- QSAR models could predict logK well, and the best QSAR model was obtained using the least-squares support vector machine method with correlation coefficient (R), leave-one/ten-out validation correlation coefficient (QLOO and QLTO) values of 0.9587, 0.8775, and 0.8732, respectively. Comparison with other methods suggested that our method performed better for predicting the logK values of β-CD complexes in terms of both accuracy and stability, especially for the discrimination of isomer structures. The results of this and previous studies demonstrate that it might be possible to use the norm index-based model to predict not only the basic physical-chemical properties, but also the chemical reaction-related constants of organic compounds.

Key words: β-cyclodextrin, Norm index, Complex stability constant, Ab initio method, Structure-activity relationship