Prediction of drug solubility in mixed solvents using computed Abraham parameters

Solubilization of drugs/drug candidates is still a challenging area in drug discovery and development investigations which computational methods could play a significant role in solving the problem. An extended version of the Jouyban-Acree model using computational Abraham's solute parameters was proposed to compute the solubility of drugs in aqueous mixtures of the cosolvents. A generally trained version of the combined model was proposed which is able to predict the solubility of drugs in water-cosolvent mixtures at various temperatures using the solubility data in neat cosolvent and water as experimental input values. The overall mean percentage deviation was 42.4 ± 59.5% which was significantly less than the overall mean percentage deviation of the similar one, i.e. log-linear model, 77.7 ± 80.8%. More general and accurate models were provided for predicting the drug solubility in water-cosolvent mixtures at various temperatures which could be employed in early stages of drug discovery and development processes to find the most appropriate cosolvent and its best composition for solubilizing a desired amount of the drug/drug candidate in a given volume of the solution.