Theoretical Prediction Model for the Acute Toxicity of Aromatic Hydrocarbons and Pesticides Compounds to Selenastrum capricornutum

Aromatic hydrocarbons and pesticides are important chemical raw materials and can cause serious pollution to environment. To study the acute toxicity of these compounds to Selenastrum capricornutum, a quantitative structure-toxicity correlation model was established. Moreover, a novel structure parameter of zero-valent atomic index (⁰B) was proposed based on the molecular topology theory, and the electrical and spatial structure properties of atoms in molecules. The electrical topological state index (E_m) and electric distance vector (M_n) of the aromatic hydrocarbons and pesticides compounds were calculated and optimized. Eight indexes of ⁰B, E₇, E₉, E₁₆, E₄₃, M₂₅, M₂₆, and M₃₆ were screened. Finally, a neural network prediction model with three-layer (8-3-1) structure was constructed. The correlation coefficient of this model is 0.9447. The mean error between the predicted value and the experimental value is only 0.224. These results show that there is a good non-linear relationship between the acute toxicity of the aromatic hydrocarbons and pesticides compounds and the zero-valent atomic index, electrical topological state index and electric distance vector. The atomic electric structure and the group fragments of =CH, =C<, =O and X (halogen groups) are the main factors affecting the acute toxicity of aromatic hydrocarbons and pesticides to Selenastrum capricornutum.