Estimating Absolute Protein-Protein Binding Free Energies by a Super Learner Model

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Biotechnology

Background:

• Protein-protein interactions are fundamental to biological processes and biotechnological applications.
• Accurate prediction of binding strength is crucial for understanding molecular mechanisms and controlling them.
• Existing computational methods for predicting binding free energy can be computationally demanding or lack sufficient accuracy.

Purpose of the Study:

• To develop a highly accurate and computationally efficient method for predicting protein-protein binding free energies.
• To create a user-friendly computational tool for estimating binding affinities from protein structures.

Main Methods:

• An ensemble machine learning model was developed using Rosetta-based quantities.
• The model was trained and validated to predict absolute binding free energies.
• The methodology was implemented as a Python pipeline named Protein Binding Energy Estimator (PBEE).

Main Results:

• The developed machine learning ensemble model achieved high accuracy in predicting binding free energies.
• PBEE demonstrated performance comparable to computationally intensive methods and superior to existing ML/DL tools.
• The pipeline enables rapid calculation of binding free energies directly from Protein Data Bank (PDB) coordinates.

Conclusions:

• The Protein Binding Energy Estimator (PBEE) provides a rapid and accurate method for predicting protein-protein binding strengths.
• This tool can significantly advance research in basic biochemistry, drug discovery, and protein engineering.
• PBEE facilitates a deeper understanding and control of protein-protein interactions in both biological systems and engineered applications.