QSAR model based on SMILES of inhibitory rate of 2, 3-diarylpropenoic acids on AKR1C3

• Computational study of 2, 3-diarylpropenoic acids as AKR1C3 inhibitors
• A ROBUST QSAR model with high R2 has been built up.
• Descriptors based on SMILES were used to build up the QSAR model.

In order to study inhibitory effect of 2, 3-diarylpropenoic acids on AKR1C3 (Aldo-keto reductase family 1 member C3), a new QSAR model based on SMILES has been built up. The descriptors of 42 compounds were calculated by CORAL software, and the Monte Carlo method was used to get the best model. The robustness of this model has been tested by Y-scrambling, Leave-one-out (LOO) cross validation and bootstrap resampling. The satisfaction model (n = 32, R2 = 0.9394, Q2 = 0.9328, s = 7.59, F = 465 for training set; n = 10, R2 = 0.9410, Q2 = 0.9064, s = 8.98, F = 127 for test set) was constructed which showed the relationship between the molecular SMILES and the inhibitory rate of 2, 3-diarylpropenoic acids on AKR1C3. This model we obtained will be helpful to find more effective AKR1C3 inhibitors.