Prediction of drug solubility from structure

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Aqueous solubility is a critical determinant of drug bioavailability.
• Predicting solubility from molecular structure is essential for efficient drug design.
• Numerous computational approaches have been developed for this purpose.

Purpose of the Study:

• To review and assess the methodologies and accuracy of computational methods for predicting aqueous solubility.
• To identify the most effective procedures for solubility prediction.
• To evaluate the current state and future needs for improving predictive models.

Main Methods:

• Review of existing literature on computational methods for aqueous solubility prediction.
• Analysis of the methodology and reported results of various predictive models.
• Specific examination of the QikProp program's solubility prediction model.

Main Results:

• Current computational methods can predict aqueous solubility with less than 1 log unit uncertainty.
• These methods are suitable for prescreening synthetic drug candidates.
• The QikProp program offers a viable approach for solubility prediction.

Conclusions:

• Effective computational tools are available for predicting drug aqueous solubility, facilitating early-stage drug discovery.
• Further improvements in prediction accuracy necessitate a large, diverse experimental database of highly accurate solubility data.
• The development of such a database will drive progress in computational solubility prediction.