酰肼类抗结核病药物的3D-QSAR研究

Quantitative structure activity relationship studies of anti-tuberculosis drug using three-dimensional structure descriptors

采用三维全息原子场作用矢量（3D-HoVAIF）对115个酰肼类抗结核病药物进行定量构效关系（QSAR）研究．运用多元线性回归（MLR）和偏最小二乘回归（PLS）建模，同时采用内部及外部双重验证的办法对所得模型稳定性能进行了深入分析和检验．MLR建模和PLS建模的复相关系数（R^2ext）、留一法交互校验复相关系数（Qk）以及外部样本校验复相关系数（Q^2ext）分别为0．766、0．663、0．748和0．733、0．614、0．715．结果表明，3D-HoVAIF能较好表征抗结核药物分子结构信息，所建QSAR模型具有良好稳定性和预测能力，为抗结核病药物的研发提供一定的理论基础．

A newly developed three-dimensional holographic vector of atomic interaction field （3D-HoV- AIF） was used to describe the chemical structures of 115 anti-tuberculosis drugs as hydrazide deriva- tives. Here quantitative structure activity relationships （QSARs） model were built by multiple linear re- gression （MLR） and partial least square regression （PLS）. The estimation stability and generalization a- bility of the model was strictly analyzed by both internal and external validations. The correlation coeffi- cients of established MLR model and PLS model （R2cum）, leave-one-out cross-validation （Q2100）, predicted values versus experimental ones of external samples （Q2ext） were 0. 766, 0. 663, 0. 748 and 0. 733, 0. 614, 0. 715, respectively. The results indicated that QSAR models had both favorable estimation stability and good prediction capabilities. Satisfactory results showed that 3D-HoVAIF could preferably express in- formation related to biological activity of hydrazide derivatives.