Drug Theoretics and Cheminformatics Laboratory, Division of Medicinal and Pharmaceutical Chemistry, Department of Pharmaceutical Technology, Jadavpur University, Kolkata 700 032, India.
Chemosphere. 2013 Jul;92(5):600-7. doi: 10.1016/j.chemosphere.2013.03.065. Epub 2013 May 2.
Four quantitative prediction models for steady-state compartmental chemical mass concentrations (Wn,g) were obtained from structural information, physiochemical properties, degradation rate and transport coefficients of 455 diverse organic chemicals using chemometric tools in a quantitative structure-fate relationship (QSFR) study. The mass ratio assessment of environmentally prevalent organic chemicals may be helpful to predict their toxicological fate in the ecosystems. Four sets of mass ratios [(1) log(Wair) from water emissions (water to air compartment), (2) log(Wair) from air emissions (within different zones of the air compartment), (3) log(Wwater) from water emissions (within different zones of the water compartment) and (4) log(Wwater) from air emissions (air to water compartment)] have been used. The developed models using genetic function approximation followed by multiple linear regression (GFA-MLR) and subsequent partial least squares (PLS) treatment identify only four descriptors for log(Wair) from water emission, six descriptors for log(Wair) from air emission, five descriptors for log(Wwater) from water emission and seven descriptors for log(Wwater) from air emission for predicting efficiently a large number of test set chemicals (ntest=182). The conclusive models suggest that descriptors such as partition coefficients (Kaw, Kow and Ksw), degradation parameters (Ksoil,Kwater and Kair), vapor pressure (Pv), diffusivity (Dwater), spatial descriptors (Jurs-WNSA-1, Jurs-WNSA-2, Jurs-WPSA-3, Jurs-FNSA-3 and Density), thermodynamic descriptors (MolRef and AlogP98), electrotopological state indices (S_dsN, S_ssNH and S_dsCH) are important for predicting the chemical mass ratios. The developed models may be applicable in toxicological fate prediction of diverse chemicals in the ecosystems.
从结构信息、物理化学性质、降解速率和运输系数四个方面，使用化学计量学工具，对 455 种不同的有机化学品进行了定量结构-命运关系（QSFR）研究，得到了 4 个用于稳态隔室化学质量浓度（Wn，g）的定量预测模型。环境中普遍存在的有机化学品的质量比例评估可能有助于预测它们在生态系统中的毒理学命运。本研究使用了四套质量比例[（1）从水排放（水到空气隔室）的 log（Wair），（2）从空气排放（空气隔室不同区域内）的 log（Wair），（3）从水排放（水隔室不同区域内）的 log（Wwater）和（4）从空气排放（空气到水隔室）的 log（Wwater）]。通过遗传函数逼近（GFA） followed 多线性回归（MLR）和随后的偏最小二乘法（PLS）处理建立的模型，仅为从水排放的 log（Wair）识别了四个描述符，为从空气排放的 log（Wair）识别了六个描述符，为从水排放的 log（Wwater）识别了五个描述符，为从空气排放的 log（Wwater）识别了七个描述符，从而可以有效地预测大量的测试集化学品（ntest=182）。结论模型表明，分配系数（Kaw、Kow 和 Ksw）、降解参数（Ksoil、Kwater 和 Kair）、蒸气压（Pv）、扩散系数（Dwater）、空间描述符（Jurs-WNSA-1、Jurs-WNSA-2、Jurs-WPSA-3、Jurs-FNSA-3 和 Density）、热力学描述符（MolRef 和 AlogP98）、电拓扑状态指数（S_dsN、S_ssNH 和 S_dsCH）等描述符对于预测化学质量比例非常重要。所开发的模型可能适用于预测生态系统中多种化学品的毒理学命运。
