Solubility prediction of drugs in methanol + water mixtures using minimum number of data points: Comparison of various models

ABSTRACT

This work reports the ability of cosolvency and activity coefficient models in the prediction of drugs solubility in methanol + water mixtures by choosing a minimum number of experimental data point. For this purpose, the available solubility datasets of drugs were gathered from papers published from 2012 to 2022 and selected seven number of each dataset as the training data whereas the rest of datapoints were the predictive data. Moreover, the effect number of model’s parameters on their accuracies was investigated by training the models with selecting of 12 data points. The results indicated that the accuracy of cosolvency models was better than the activity coefficient models; so that, the simple cosolvency model of Jouyban-Acree which does not require the physicochemical properties of drugs including melting point, fusion enthalpy and also molecular parameters of van der Waals area and volume predicted the drugs solubility with more accuracy (MRDs% of 4.1).

KEYWORDS:

• Solubility prediction
• aqueous binary mixtures of methanol
• minimum experiments
• cosolvency models
• activity coefficient
• interpolation

Acknowledgments

The authors wish to thank financial support from the graduate council of the Tabriz University of Medical Science.

Disclosure statement

No potential conflict of interest was reported by the authors.

Credit authorship contribution statement

Parisa Jafari: Conceptualization, Supervision, Investigation, Writing-review & editing, Formal analysis, original draft.

Elaheh Rahimpour: Writing-review & editing, Investigation

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00319104.2023.2289180