Localization bond index was used for acute toxicity of oxygenated polycyclic aromatic hydrocarbons to zebrafish embryos
Oxygenated polycyclic aromatic hydrocarbons(OPAHs)were persistent organic pollutants,containing one or more base oxygens on the benzene ring of PAHs.OPAHs were more toxic,carcinogenic,teratogenic,and mutagenic than parental PAHs.Therefore,it was very important to explore efficient research methods to establish toxicity database.In order to investigate the quantitative structure-toxicity relationship(QSTR)between the acute toxicity of OPAHs to zebrafish embryos and their molecular structures,a novel structural index,the localization bond index L,was derived based on the spatial topology of atoms in OPAHs molecules.Moreover,the electrical topological state indices Em,electrical distance vectors Mn were calculated for 32 molecules of OPAHs compound,and the electrical topological state index E16,the electrical distance vectors M14,M21and M32 were selected as structural descriptors.The electrical topological state index and electrical distance vectors were combined with localization bond index L,and they were introduced to the regression analysis of the acute toxicity of OPAHs to zebrafish embryos.Using the five molecular structural indices as enter the variable point value,and the neural network structure was performed using 5-3-1,a neural network model which predicted the acute toxicity lgEC50 of OPAHs was constructed.The total correlation coefficient RT was 0.9826,and the average error between the experimental and the predicted values of acute toxicity was only 0.112.The results showed that the acute toxicity of OPAHs had a good nonlinear relationship with the structural parameters of the localization bond index,the electrical topological state index,and the electrical distance vectors.The study provided theoretical guidance for environmental pollution and ecological risk assessment of OPAHs.
oxygenated polycyclic aromatic hydrocarbonslocalization bond indexzebrafish embryosquantitative structure-toxicity relationship(QSTR)neural network
NETL
NSTL
万方数据
