手性羟酸和氨基酸类化合物的构效关系研究-上海有机化学研究所.PDFVIP

2009 年第 67 卷 化 学 学 报 Vol. 67, 2009 第 2 期, 145～150 ACTA CHIMICA SINICA No. 2, 145～150 ·研究论文· 手性羟酸和氨基酸类化合物的构效关系研究 刘 东 a,b 章文军 a 许 禄*,b (a 河北工业大学 应用化学系 天津 300130) (b 中国科学院长春应用化学研究所 长春 130022) 摘要 构造了用于手性参数衍生的距离矩阵和扩展的距离矩阵, 并且观察了与羟酸类和氨基酸类化合物手性薄层色谱 保留指数的相关性, 得到了具有启示性的结果. 引进了扩展 eAm 指数和连乘指数, 以及综合运用了三种算法. 所得参 数进行了多元回归分析, 得到了单一变量和变量组合的分析结果. 与多元回归分析相比较, 运用人工神经网络法, 使 结果得到了较大的改善. 所引进的方法对于化合物构效关系研究具有普遍可用性. 关键词 手性化合物; 构效关系研究; 手性指数; 回归分析; 人工神经网络 Quantitative Structure-Activity/Property Relationships for Chiral Hydroxy Acids and Amino Acids Liu, Donga,b Zhang, Wenjuna Xu, Lu*,b a ( Hebei University of Technology, Department of Applied Chemistry , Tianjin 300130) (b Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022) Abstract To study quantitative structure-activity/property relationships (QSAR/QSPR) for hydroxy acids and amino acids, the distance matrices and the augmented distance matrices have been constructed to derive the chiral parameters. Several regression results obtained by using these parameters have been achieved. Chiral topological indices eAm by extending Am indices, chiral product parameters as well as their combina- tions were used to explain the variance in thin-layer chromatographic indices. For the same purpose, artifi- cial neural networks were applied, and the results were i