Predicting aqueous solubility from structure

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Aqueous solubility is a critical physicochemical property influencing drug absorption, distribution, metabolism, and excretion (ADME).
• Drug solubility impacts the efficiency and success of high-throughput screening (HTS) campaigns.
• Accurate solubility prediction is vital for early-stage drug development.

Purpose of the Study:

• To critically review methods for predicting drug aqueous solubility.
• To discuss challenges and limitations in current solubility prediction techniques.
• To highlight the need for improved prediction models for complex systems.

Main Methods:

• Survey of various computational and experimental methods for solubility prediction.
• Analysis of factors affecting the applicability and accuracy of different techniques.
• Discussion of trade-offs between speed, accuracy, and transparency of prediction methods.

Main Results:

• Current computational programs can provide solubility estimates within an order of magnitude in favorable cases.
• Various methods exist, each with specific strengths and weaknesses.
• Significant challenges persist for predicting solubility in non-ideal conditions.

Conclusions:

• While progress has been made, accurate aqueous solubility prediction remains an active area of research.
• Further development is needed for models predicting solubility in challenging solvents like DMSO or for charged solutes.
• Balancing prediction speed, accuracy, and interpretability is key for practical applications.