Docking to multiple pockets or ligand fields for screening, activity prediction and scaffold hopping.

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• False-negatives in activity prediction pose challenges in drug discovery.
• Recent advances include multi-conformational docking (mPockDock) and atomic property field docking (mApfDock).

Purpose of the Study:

• To compare the efficacy of mPockDock and mApfDock in predicting compound activity.
• To evaluate their performance in scaffold hopping applications.

Main Methods:

• Utilized multi-conformational docking (mPockDock) for flexible compound models.
• Employed atomic property field docking (mApfDock) using co-crystallized ligand-derived fields.
• Assessed performance using Area Under the Curve (AUC) metrics.

Main Results:

• mApfDock achieved an AUC of 90.4%, outperforming mPockDock's 83.8%.
• mApfDock showed superior performance for compounds requiring significant induced-fit pocket adaptations.
• mPockDock was more effective when co-crystallized ligands lacked chemical and binding diversity.

Conclusions:

• Both mPockDock and mApfDock are efficient for scaffold hopping.
• These methods are complementary when co-crystallized complex coverage is limited.
• Their performance converges when complex diversity is sufficient.