摘要
Quantitative structure activity relationship （QSAR） studies were performed on 45 anthranilic acid derivatives for their potent allosteric inhibition activities of HCV NSSB polymerase. Genetic algorithm based genetic function approximation （GFA） method of variable selection was used to generate the model. Highly statistically significant model with r^2 = 0.966 and r^2cv = 0.951 was obtained when the number of descriptors in the equation was set to 5. High r^2pred value of 0.884 indicates the good predictive power of the best model. Spatial descriptors of radius of gyration （RadOfGration）, molecular volume （Vm）, length of molecule in the z dimension （Shadow-Zlength）, thermodynamic descriptors of the octanol/water partition coefficient （LogP） and molecular refractivity index （MR） showed enormous contributions to HCV NS5B polymerase inhibition. The validation of the model was done by leave-one-out （LOO） test, randomization tests and external test set prediction. The model gives insight on indispensable structural requirements for the activity and can be used to design more potent analogs against HCV NSSB polymerase.
Quantitative structure activity relationship （QSAR） studies were performed on 45 anthranilic acid derivatives for their potent allosteric inhibition activities of HCV NSSB polymerase. Genetic algorithm based genetic function approximation （GFA） method of variable selection was used to generate the model. Highly statistically significant model with r^2 = 0.966 and r^2cv = 0.951 was obtained when the number of descriptors in the equation was set to 5. High r^2pred value of 0.884 indicates the good predictive power of the best model. Spatial descriptors of radius of gyration （RadOfGration）, molecular volume （Vm）, length of molecule in the z dimension （Shadow-Zlength）, thermodynamic descriptors of the octanol/water partition coefficient （LogP） and molecular refractivity index （MR） showed enormous contributions to HCV NS5B polymerase inhibition. The validation of the model was done by leave-one-out （LOO） test, randomization tests and external test set prediction. The model gives insight on indispensable structural requirements for the activity and can be used to design more potent analogs against HCV NSSB polymerase.
基金
supported by the National Natural Science Foundation of China (No. 30500339)
Natural Science Foundation of Zhejiang Province (NO.Y407308)
the Sprout Talented Project Program of Zhejiang Province (No. 2008R40G2020019)
