Study On Structure-Toxicity Relationship Of Substituted Phenols And Anilines Towards Algae

With the development of society,progress of industry together with human activities,a lot of chemical substances were released into the environment.Phenol,aniline and their derivatives,are widely used in industrial materials.However,these chemicals prone to contaminate the surface waters or the ground waters through drift of aerial spray and/or watershed drainage,posing a serious threat to the environment and human health.Ecologically,algae play a significant role in the aquatic ecosystem as a dominant producer,providing energy and oxygen at higher levels.Adverse effects of these toxic chemicals to algae may reduce the primary productivity of the ecosystem and further disrupt the food web.The aquatic toxicity data of these substances to aquatic ecosystem are required in order to evaluate their hazard and risk.Experimental methods of aquatic toxicity data are both time-consuming and expensive.The quantitative structure-activity relationship(QSAR)could be an alternative method to solve these problems.In this study,quantitative structure-toxicity relationship(QSTR)models with the same mathematical structure were proposed for predicting the multiple toxicity endpoints of substituted phenols and anilines towards algae based on norm indexes.This study has carried out from the following aspects,the research results are:Firstly,study on the toxicity of phenols and anilines to Chlorella vulgaris.A multiple toxicity endpoint-structure modeling study was performed with uniform norm-index descriptors for predicting the aquatic toxicity of organic chemicals towards C.vulgaris.The dataset for this work included the 96-h algal toxicity values of 67 substituted phenols and anilines to C.vulgaris.R2 values of 0.897,0.876,0.848 and 0.860 and Q2 values of 0.854,0.831,0.785 and 0.794 for pIC50,pIC20,pLOEC and pNOEC,respectively.These satisfactory results obtained in this work suggested that it might be possible to develop QSTR models with the same mathematical structure for predicting multiple toxicity endpoints successfully via norm index descriptors.Secondly,study on the toxicity of phenols to Dunaliella tertiolecta.Based on the same norm-index descriptors,three models based on the same descriptors were established to predict the multiple toxicity endpoints of 30 substituted phenols and anilines to Dunaliella tertiolecta.R2 values are 0.931,0.940 and 0.943,respectively.Q2 values are 0.879,0.897 and 0.902,respectively.These satisfactory results showed that it is reliable to develop QSTR models with the same descriptors for predicting the toxicities of phenols to Dunaliella tertiolecta.Thirdly,study on the toxicity of phenols and anilines to Pseudokirchneriella subcapitata.A model was built based on the norm index to predict the toxicity of 58 phenols towards Pseudokirchneriella subcapitata.For the built model,R2 is 0.866,Q2 is 0.840.Results show that the model based on the norm index could be used to predict the toxicity of organic pollutants towards Pseudokirchneriella subcapitata.In this study,the OSTR models were built to predict the toxicities of organic pollutants to algae based on the norm indexes proposed by our group.Leave-one-out cross validation,Y-randomized validation,external validation,application domain analysis and comparison with the reference models demonstrated the accuracy,robustness and reliability of these models.The results indicated that the developed models could produce satisfactory predictive results for single and multiple toxicity endpoints.Results in this study showed that the descriptors were effect.The norm indexes are purely derived from chemical structures therefore,the norm indexes descriptors were accuracy and universality and generality.