Traditional salt screening, while crucial for enhancing API solubility and stability, is often unpredictable and resource-intensive due to extensive experimental testing. Faced with the challenge of a difficult-to-form salt compound, the XtalPi team leveraged our virtual and experimental screening workflow to identify an optimal salt that met formulation development requirements.
In this case, we demonstrated the capabilities of our workflow:
- Our Salt Formation Propensity Prediction model, a hierarchical AI-based system, screened 4,070 potential experiments to recommend 40 high-probability salt-forming counterions and solvent combinations based on molecular and interaction features.
- Experimental screening and characterization using model recommendations—selected for their high salt formation probability with the top 10 counterions—resulted in the identification of seven distinct crystalline salt forms with five different counterions, each demonstrating high stability and salt formation propensity as predicted
- Solubility and hygroscopicity evaluations identified one lead salt form with superior physicochemical properties, meeting solubility and stability requirements and making it ideal for further development.
