Combining MVD and Ridge Method to Predict CDK2 Inhibition.

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Molegro Virtual Docker (MVD) is a widely used docking program for protein-ligand interactions.
• MVD offers flexibility through 16 combinations of search algorithms and scoring functions.
• Docking results from MVD have been successfully applied to predict protein inhibition.

Purpose of the Study:

• To integrate MVD with Scikit-Learn's Ridge regression for enhanced predictive modeling.
• To explore the scoring function space for improved computational drug design.
• To predict the inhibition of cyclin-dependent kinase 2 (CDK2) using this integrated approach.

Main Methods:

• Utilized Molegro Virtual Docker (MVD) for docking simulations.
• Integrated MVD outputs with Scikit-Learn's Ridge regression machine learning model.
• Applied the combined method to predict cyclin-dependent kinase 2 (CDK2) inhibition.

Main Results:

• The integrated MVD and Scikit-Learn model demonstrated superior predictive performance.
• The computational model achieved better binding affinity prediction compared to classical scoring functions.
• The study explored the concept of scoring function space for model development.

Conclusions:

• Combining MVD with machine learning, specifically Ridge regression, offers a powerful approach for predicting protein inhibition.
• This integrated method provides enhanced predictive accuracy for binding affinity.
• The developed computational model shows potential for drug discovery and development efforts.