POSIT: Flexible Shape-Guided Docking For Pose Prediction.

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Drug discovery

Background:

• Structure-based drug design (SBDD) relies on accurate prediction of ligand poses within protein targets.
• Existing pose prediction methods face challenges related to data quality and availability.

Purpose of the Study:

• To introduce and validate a novel SBDD approach (POSIT) for improved pose prediction.
• To assess POSIT's performance against existing methods using retrospective and prospective analyses.
• To evaluate POSIT's utility in guiding industrial drug discovery projects, particularly lead optimization.

Main Methods:

• Development of the POSIT framework, emphasizing the link between pose prediction and experimental structural data quality.
• Retrospective analysis using three diverse datasets to benchmark POSIT against other methods.
• Prospective validation over 2.5 years across multiple industrial SBDD projects, involving 71 crystal structures.

Main Results:

• POSIT demonstrates strong performance in pose prediction, outperforming existing methods in specific scenarios.
• The framework effectively guides decision-making in lead optimization campaigns.
• Prospective use shows POSIT's practical applicability and impact in real-world drug discovery.

Conclusions:

• POSIT offers a robust and effective approach for structure-based drug design and lead optimization.
• The study presents the largest prospective validation of a pose prediction method to date.
• POSIT provides valuable insights into the impact of computational tools on prospective drug design.