Developing a Naïve Bayesian Classification Model with PI3Kγ structural features for virtual screening against PI3Kγ:...

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Immunology

Background:

• Phosphatidylinositol 3-kinase gamma (PI3Kγ) is a crucial target for immune signaling modulation.
• Developing selective PI3Kγ inhibitors is challenging due to the conserved ATP-binding pocket across kinase subtypes.

Purpose of the Study:

• To develop a virtual screening strategy for identifying novel, selective PI3Kγ inhibitors.
• To improve the accuracy and reliability of virtual screening methods for drug discovery.

Main Methods:

• A Naïve Bayesian Classification (NBC) model was developed, integrating molecular docking and pharmacophore analysis.
• The model utilized multiple PI3Kγ conformations and was validated using internal cross-validation and external prediction.
• Virtual screening was performed on an analog dataset based on a reference compound (JN-PK1).

Main Results:

• The integrated NBC model demonstrated significantly improved enrichment of active PI3Kγ inhibitors compared to single-conformation methods.
• The virtual screening successfully identified a novel PI3Kγ inhibitor with higher activity than the reference compound.
• Biological assays confirmed the accuracy and reliability of the developed NBC model.

Conclusions:

• The developed virtual screening strategy, combining molecular docking, pharmacophore, and NBC, is effective for discovering selective PI3Kγ inhibitors.
• This approach offers valuable guidance for future drug discovery efforts targeting PI3Kγ.
• The identified novel inhibitor warrants further investigation for therapeutic potential.