Enhancing PI3Kγ inhibitor discovery: a machine learning-based virtual screening approach integrating pharmacophores,...

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Drug Discovery

Background:

• Phosphoinositide 3-kinase gamma (PI3Kγ) is a lipid kinase predominantly in leukocytes.
• PI3Kγ is implicated in critical pathways for tumors, inflammation, and autoimmune diseases.
• Targeting PI3Kγ with inhibitors is a key area in pharmaceutical research.

Purpose of the Study:

• To develop and validate a machine learning model for virtual screening of PI3Kγ inhibitors.
• To identify novel chemical entities with potential PI3Kγ inhibitory activity.
• To provide insights into favorable structural fragments for PI3Kγ inhibitor design.

Main Methods:

• Development of a Naïve Bayesian Classification (NBC) model integrating molecular descriptors, fingerprints, docking, and pharmacophore data.
• Virtual screening of the ChEMBL database using the optimized NBC model.
• In silico validation of the model's predictive performance.

Main Results:

• The optimal NBC model demonstrated high accuracy in distinguishing active from inactive PI3Kγ compounds.
• The model successfully identified potential novel PI3Kγ inhibitors from the ChEMBL database.
• Key molecular fragments contributing to PI3Kγ inhibition were elucidated.

Conclusions:

• Machine learning, specifically the developed NBC model, offers a powerful strategy for accelerating PI3Kγ inhibitor discovery.
• The validated model can guide the rational design of next-generation PI3Kγ-targeted therapeutics.
• This approach provides valuable insights for developing treatments for PI3Kγ-related diseases.