Study On Prediction Methods Of Toxicity Of Organic Compounds To Aquatic Organisms

With the continuous development of heavy industry and aquaculture,a large amount of organic pollutants enter the water environment,causing serious harm to human health and aquatic systems.Currently,machine learning methods can efficiently process data and build quality models.Ensemble learning is built on multiple base classifiers,which typically have better predictive power than the performance of any single model.Therefore,based on the experimental data collected by literature collection,this study uses machine learning and ensemble learning methods to predict the bioconcentration factors of organic compounds and aquatic acute toxicity,and then analyze the toxicity mechanism of organic compounds on aquatic organisms.Bioconcentration factors and median lethal concentrations(LC₅₀s)are important when assessing risks posed by organic pollutants to aquatic ecosystems.Various quantitative structure–activity relationship models have been developed to predict bioconcentration factors and classify acute toxicity of aquatic organisms.In the establishment of Bioconcentration factors prediction model,two regression models were developed by using Recursive Feature Elimination method combined with Support Vector Machine and Multiple Linear Regression algorithms.We calculated2D molecular descriptors from a data set containing 500 diverse chemicals in our regression model.In the classification prediction of aquatic acute toxicity,we built three ensemble models using three machine learning algorithms and calculated 12molecular fingerprints from a dataset containing 400 diverse chemicals in our classification models.In the regression model,the RFE-SVM model presented better predictive performance.The R² and_??were 0.860 and 0.757,respectively.Other indicators could also indicate that the regression model made good predictions and could efficiently predict a new set of compounds following standards set by Golbraikh,Tropsha,and Roy.In the classification models,the ensemble-SVM classification model gave an overall accuracy,sensitivity,specificity,and AUC?area under the receiver operating characteristic curve?of 92.2,95.1,86.0,and 0.965,respectively,in a five-fold cross-validation and of 87.3,92.6,76.0,and 0.940,respectively,in an external validation.These parameters indicated that our ensemble-SVM model was more stable and gave more accurate predictions than previous models.The model could therefore be used to effectively predict aquatic toxicity and assess risks posed to aquatic ecosystems.Moreover,we identified several chemical structures associated with bioconcentration factors and aquatic acute toxicity by analyzing two models,in particular the structure aaCH,aromatic structures,hydrogen bonding groups and water partition coefficients,in future aquatic toxicology experiments.It should be given more attention in the risk assessment of aquatic ecosystems.In summary,this paper had the following innovative work:?1?establishing a QSAR classification model for acute aquatic toxicity and better performance parameters were obtained;?2?Aquatic biological toxicity was analyzed by combining bioconcentration factors with acute toxicity.