ProBAPred: Inferring protein-protein binding affinity by incorporating protein sequence and structural features

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Protein-protein interactions (PPIs) are crucial for biological processes, cellular pathways, and diseases.
• Understanding PPIs aids in constructing interaction networks and deciphering protein functions.
• Limited computational methods exist for accurately estimating protein-protein binding free energy.

Purpose of the Study:

• To develop a novel ensemble computational framework, ProBAPred (Protein-protein Binding Affinity Predictor), for quantitative estimation of protein-protein binding affinity.
• To identify and utilize informative sequence and structural features for predicting binding affinity.
• To provide accurate real-value regression models for characterizing protein-protein binding affinity.

Main Methods:

• Collected and calculated a comprehensive set of sequence and structural features from protein binding complex datasets and literature.
• Employed feature selection using the WEKA package to identify the most informative features.
• Developed an ensemble computational framework (ProBAPred) for predicting protein-protein binding affinity.

Main Results:

• ProBAPred achieved a Mean Absolute Error (MAE) of 1.657 kcal/mol on an independent test set.
• The ensemble method demonstrated a high correlation coefficient, outperforming existing methods.
• Identified key sequence and structural features contributing to binding affinity prediction.

Conclusions:

• ProBAPred offers a robust computational approach for estimating protein-protein binding affinity.
• The developed regression models can significantly aid in the computational characterization of protein binding.
• ProBAPred facilitates experimental studies by providing accurate affinity predictions.