Luo-nan Xu, Gui-Zhao Liang, Zhi-liang Li, Juan Wang, Peng Zhou
College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400044, China.
J Mol Graph Model. 2008 Jun;26(8):1252-8. doi: 10.1016/j.jmgm.2007.11.007. Epub 2007 Dec 4.
A newly developed descriptor, three-dimensional holographic vector of atomic interaction field (3D-HoVAIF), is used to describe the chemical structures of Aza-bioisosteres of anthrapyrazoles (Aza-APs). 3D-HoVAIF contains three non-bonded (electrostatic, van der Waals and hydrophobic) factors, which are directly related to bioactivities and utilized to express intramolecular potential energies, employed to several different models on 35 anticancer agents Aza-APs. After variable screening by stepwise multiple regression (SMR) technique, a partial least square (PLS) regression model is built with 3D-HoVAIF. The model is satisfactory comparing to reference since correlation coefficients of molecular modeling (R(2)), cross-validation (q(2)) and standard deviation of estimation (S.D.) are 0.865, 0.722 and 0.299, respectively, showing that the model has favorable estimation and prediction capabilities. It is illustrated that information related to the anticancer activity of Aza-APs could preferably be expressed by 3D-HoVAIF with definite physicochemical meanings and easy structural interpretation for Aza-APs. 3D-HoVAIF is proved to be potent in relation with bioactivities, while overcoming many demerits of traditional 3D structural characterization methods.
一种新开发的描述符——原子相互作用场的三维全息向量(3D-HoVAIF),用于描述蒽并吡唑类氮杂生物电子等排体(Aza-APs)的化学结构。3D-HoVAIF包含三个非键合(静电、范德华和疏水)因子,这些因子与生物活性直接相关,并用于表达分子内势能,应用于35种抗癌药物Aza-APs的几种不同模型。通过逐步多元回归(SMR)技术进行变量筛选后,用3D-HoVAIF建立了偏最小二乘(PLS)回归模型。与参考文献相比,该模型令人满意,因为分子建模相关系数(R(2))、交叉验证(q(2))和估计标准差(S.D.)分别为0.865、0.722和0.299,表明该模型具有良好的估计和预测能力。结果表明,可以用具有明确物理化学意义且易于对Aza-APs进行结构解释的3D-HoVAIF较好地表达与Aza-APs抗癌活性相关的信息。3D-HoVAIF被证明与生物活性相关,同时克服了传统三维结构表征方法的许多缺点。