信息理论指数及其在有机磷农药毒性预测研究中的应用

Information Theoretic indices and Application to Predicting the Toxicities of Organophesphorus Pesticides

为探讨化合物结构与毒性关系研究方法，将定量描述分子结构特征的信息理论指数、价分子连接性指数应用于有机磷农药毒性预测研究，提出了基团信息参数。运用判别分析、多元非线性回归分析，依据１１４种有机磷农药急性毒性数据资料，按动物不同给药途径，分别建立定性、定量预测模型，定性判别回代符合率均在８３．３３％以上，定量回归方程相关指数Ｒ２均在０．８１以上，回归残差呈正态分布。预测分析，预测值均在ＬＤ５０测定值的９５％可信区间内。预测效果满意。

he information theoretic indices and the valenca molecular connectivity indices as the characteristics of molecular structure were applied to predicting the toxicities of organophosphorus pesticides.Radical information parameter was suggested.According to the data on acute toxicities Of 114 organophosphorus Pesticides and different channels that the pesticides were given, the qualitative and quantitative prediction models were estabilished along with a discriminatory analysis and a multiple non-linear regression analysis. Rates of qualitative discriminatory analysis were not less than 83. 33%. The correlation indices of quantitative regression models (R2) were over 0. 81. Residuals showed the normal distribution. Predicted values LD50 by the regression equations were within a 95% confidence interval of observed values.