Chem. J. Chinese Universities ›› 1997, Vol. 18 ›› Issue (1): 60.

• Articles • Previous Articles     Next Articles

Using Molecular Graphics, Molecular Mechanics, Quantum Chemistry and Electrostatic Potential Methods to Study Structure-Property Relationship on Pesticides(Ⅸ)─Effect of Selecting Structural Parameters and Calculating Methods on Improving the Acc

WANG Xia1, YUAN Man-Xue1, LAI Cheng-Ming1, LIU Jie2, LI Zheng-Ming2   

  1. 1. National Key Laborafory of Element-Organic Chemistry, Department of Chemistry, Nankai University, Tianjin, 300071;
    2. Elemento-Organic Chemistry Institute, Nankai University
  • Received:1996-04-24 Online:1997-01-24 Published:1997-01-24

Abstract: In this paper, we discussed mainly two sldes of factors-structural parameters and calculating methods, which improved the accuracy of forecasting activities on sulfony-lure a molecules.When selecting the structural parameters, we considered not only the radi-cal effect of molecules, but also their characteristics as a whole-the values of conforma-tion difference between the most active molecule and other molecules were computed, and at the same time, we applied the optimum calculating method - Artificial Neural Networks (ANN).Then we calculated the PI50 values of 35 newly synthesized sulfonylurea molecules,which were almost in agreement with experimental data(mse=0.073) and superior to those of multiregression analysis (mse=0.82).

Key words: Sulfonylurea herbicides, Activity, Structural parameters, Artificial neural networks

TrendMD: