【摘要】 在环境水体中,可电离有机化合物(IOCs)可解离为分子和离子形态。研究表明,IOCs离子形态的环境行为、毒性效应等都与其分子形态存在较大差异,因而在研究IOCs环境行为、毒性效应时不应忽略离子化的影响。在构建IOCs相关预测模型时如何表征离子化的影响是当前研究的重要内容之一。探讨了采用基于形态修正的描述符构建IOCs水生毒性预测模型的可行性。具体而言,采用逐步多元线性回归(MLR)方法,构建了可预测63种取代酚、取代苯甲酸和取代苯胺等IOCs对大型溞急性毒性的定量结构-活性关系(QSAR)模型。与仅采用分子形态描述符的模型相比,使用基于形态修正描述符的模型决定系数(R2)、去一法交叉验证系数(Q2LOO)、外部验证系数(Q2EXT)等参数从0.622～0.705提高到了0.840～0.875,表明基于形态修正描述符的模型具有更好的拟合优度、稳健性和预测能力。因此,在将来的研究中,可采用基于形态修正的描述符构建IOCs水生毒性效应预测模型。更多还原

【Abstract】 Ionogenic organic chemicals(IOCs) may ionize to form anion and/or cation in aquatic environment. It had been elucidated that the environmental behavior,toxic effects of ionic form differ greatly from that of neutral form. Thus,ionization is nonnegligible in performing the research of the environmental behavior,toxic effects of IOCs. To date,how to characterize the ionization is one of the critical issues in developing the predictive models for IOCs. Here,the feasibility of using chemical form adjusted descriptors as predictive variable to derive model for the endpoint of aquatic toxicity was studied. In this regard,the acute toxicity data of 63 substituted phenols,anilines and benzoic acids to Daphnia magna were collected firstly. Then,the quantitative structure-activity relationship(QSAR) model was developed by stepwise multiple linear regressions(MLR) analysis. The modeling results indicated that the values of coefficient determination(R2),cross validated Q2 leave-one-out(Q2 LOO),external validation coefficient(Q2 EXT) for the QSAR model based on chemical form adjusted descriptors was significantly improved from 0.622-0.705 to 0.840-0.875,compared with that of the model constructed from neutral form descriptors only.The results indicated that the QSAR model based on chemical form adjusted descriptors had high goodness-of-fit,robustness,and predictive ability. Thus,the predictive models of IOCs for aquatic toxicity could be developed by employing chemical form adjusted descriptors in the future QSAR modeling.更多还原