The selection of organic solvents is of great importance in the process of chemical and drug development. However, the accuracy of current prediction methods is limited. This pipeline based on the DNN and lightGBM models provides accurate solubility predictions. The predictions are very general and can attempt to predict the solubility of compounds in many organic solvents at a given temperature, which allows for rapid solvent screening..
© Computational Pharmaceutics Group. 2019. All Rights Reserved.