Neural Networks with Molegro Data Modeller.

Area of Science:

• Computational chemistry
• Biophysics
• Machine learning

Background:

• Artificial neural networks are fundamental to deep learning and modeling complex systems like proteins.
• Predicting protein-ligand binding affinity is crucial for drug discovery.
• Cyclin-dependent kinase 2 (CDK2) is a key target for anticancer drug development.

Purpose of the Study:

• To develop and evaluate a neural network model for predicting the inhibition of CDK2.
• To assess the model's performance against classical scoring functions.

Main Methods:

• Utilized atomic coordinates of a CDK2-Cyclin A2 complex and BindingDB data.
• Employed Molegro Data Modeller to construct a regression model.
• Incorporated features derived from the Molegro Virtual Docker (MVD) program.

Main Results:

• The developed neural network model achieved superior predictive performance.
• The model's accuracy surpassed that of classical scoring functions in predicting CDK2 inhibition.

Conclusions:

• Neural network-based regression models can effectively predict protein-ligand binding affinity.
• This approach offers a promising tool for accelerating anticancer drug discovery targeting CDK2.