Background: Because of their intrinsic ability to induce physiological effects in humans at low doses, pharmaceuticals and personal care products (PPCPs) are a unique group of emerging environmental pollutants. A number of studies have confirmed the occurrence of different PPCPs in the environment, which raises concerns about possible adverse effects on humans and wildlife. The removal of PPCPs from wastewaters has become a major activity to reduce pollution due to their adverse effects on humans and aquatic ecosystems.

Methods: This study aimed to design a Quantitative Structure Activity Relationship (QSAR) model for the removal of 57 PPCPs from wastewater treatment plants (WWTPs) of historical data obtained from plants located in South Korea. The target compounds of PPCPs were optimised geometrically using a Forcite-Geometry code, assembled in Material Studio 2016.

Results: The removal efficiency of PPCPs is dependent on several preliminary molecular descriptors including rotatable bonds (RBs), hydrogen bond donor (HBD), total molecular mass (TMM), binding energy (BE), atom count (AC), element count (EC), total energy (TE), total dipole (TD), highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO). A Genetic Function Approximation (GFA) method was adopted to perform regression analysis and create correlation between experimental data (literature) and measured data (QSAR model).

Conclusion: A QSAR model equation was established and used to predict removal efficiency of 57 PPCPs; the results obtained showed goodness of fit, R² greater than 0.90 indicating that the internal and external validations were also performed on the model.

背景：由于药物和个人护理产品 (PPCP) 具有在低剂量下诱导人体生理效应的内在能力，因此它们是一类独特的新兴环境污染物。多项研究已证实环境中存在不同的 PPCP，这引起了人们对可能对人类和野生动物产生不利影响的担忧。由于其对人类和水生生态系统的不利影响，从废水中去除 PPCP 已成为减少污染的主要活动。

方法：本研究旨在设计一个定量结构活性关系 (QSAR) 模型，用于从来自韩国工厂的历史数据的污水处理厂 (WWTP) 中去除 57 种 PPCP。使用在 Material Studio 2016 中组装的 Forcite-Geometry 代码对 PPCP 的目标化合物进行几何优化。

结果：PPCPs的去除效率取决于几个初步的分子描述符，包括可旋转键（RBs）、氢键供体（HBD）、总分子量（TMM）、结合能（BE）、原子数（AC）、元素数（ EC)、总能量 (TE)、总偶极子 (TD)、最高占据分子轨道 (HOMO) 和最低未占据分子轨道 (LUMO)。采用遗传函数逼近 (GFA) 方法进行回归分析，并在实验数据（文献）和测量数据（QSAR 模型）之间建立相关性。

结论：建立了QSAR模型方程，用于预测57种PPCPs的去除效率；获得的结果显示拟合优度，R ²大于 0.90，表明也对模型进行了内部和外部验证。