基于QSAR模型预测有机污染物在XAD与空气中的分配系数

基于定量构效关系(QSAR),运用线性(逐步多元回归MLR)和非线性(支持向量机SVM)两种计算方法开发了两种可靠且高效预测聚苯乙烯二乙烯基苯树脂(XAD)和空气之间分配系数(K_XAD-A)的模型.构建模型的数据包含醇类(Alcohols),苯类(Benzenes),多氯联苯(PCBs)和多环芳香烃(PAHs)等,共计70种有机污染物.两个模型的决定系数R²_adj和外部验证系数Q²_ext均在0.930以上,同时所有物质均在定义的应用域内,结果表明两种QSAR模型有较高的拟合度、稳健性和较为优秀的预测能力,且非线性(SVM)模型比线性(MLR)模型的拟合效果更好.

Prediction of organic pollutions partition coefficients between XAD and air based on QSAR models

Based on quantitative structure-activity relationship (QSAR), two reliable and efficient models for predicting partition coefficients between polystyrene-divinylbenzene resin (XAD) and air (K_XAD-A) were developed by using linear (stepwise multiple regression, MLR) and nonlinear (support vector machine, SVM) methods. The data included alcohols, benzenes, polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), and a total of 70 organic pollutants. The determination coefficient (R²_adj) and external validation coefficient (Q²_ext) of the two models are all above 0.930, and substances are all in the defined application domain. The QSAR models both have good robustness and excellent prediction ability. Moreover, the fitting effect of nonlinear (SVM) model is better than linear (MLR) model.