S-型和J-型浓度-效应关系混合物毒性定量预测研究

混合物联合毒性预测主要基于浓度加和(CA)与独立作用(IA)模型。CA和IA只能评估加和作用混合物，目前缺少可靠方法定量预测具有相互作用(协同和拮抗)的混合物毒性。对于非单调J-型浓度-效应关系混合物的相互作用判断及其联合毒性预测，现有方法存在不能评估和预测难题。本项目以高精度微板毒性分析法测定多种污染物的S-型和J-型浓度-效应曲线(CRC)；基于CA和IA模型判断S-型CRC混合物的加和、协同和拮抗作用，应用线性回归和主成分分析等方法构建S-型CRC混合物相互作用预测模型；改进传统CA模型建立J-型CRC混合物加和模型，并用于判断和揭示J-型CRC混合物相互作用方式与规律；以J-型CRC混合物加和模型推导和计算J-型CRC的特征参数，提出利用支持向量机构建J-型CRC混合物毒性预测模型，并准确预测混合物联合毒性，为解决混合物毒性的定量预测与风险评估难题提供新方法。

The joint toxicity of the mixture is mainly based on the concentration addition (CA) and independent action (IA) models, which can only be used to assess the mixture with additive effect. Reliable method is lacked for the accurate prediction of the interactive mixture toxicity (synergistic and antagonistic interaction). The existing models are failed to evaluate the interaction type of the mixture with non-monotone J-shaped concentration-effect relationship, and to predict the mixture toxicity. The presented project applies the microplate toxicity analysis (MTA) method to test the S-shaped and J-shaped concentration-response curves (CRC) of the pollutants. The determinations of the additive, synergistic and antagonistic interactions of the mixture with S-shaped CRC are based on the CA and IA models. The linear regression and principal component analysis method are used to develop models that are able to predict the toxicities of the S-shaped CRC interactive mixture. The classical CA model is modified into a new additive model for mixture with J-shaped CRC, which is used to determine the interaction type of the mixture. The characteristic parameters of J-shaped CRC mixture are deducted and calculated from the new additive model. The support vector machine is used to build a new model for the mixture with J-shaped CRC, which is used to predict the joint toxicity of mixtures. The results provide new methods for quantitative prediction and risk assessment of mixture toxicity.