SSIPe: accurately estimating protein-protein binding affinity change upon mutations using evolutionary profiles in...

Area of Science:

• Computational biology
• Structural biology
• Bioinformatics

Background:

• Protein-protein interactions (PPIs) are vital for cellular functions.
• Mutations at protein interfaces can alter PPI stability, leading to diseases.
• Accurate estimation of binding affinity changes (ΔΔGbind) is essential for diagnosing genetic diseases and annotating protein functions.

Purpose of the Study:

• To develop a computational method for accurate prediction of binding affinity changes (ΔΔGbind) caused by mutations at protein interfaces.
• To improve the understanding of mutation effects on protein-protein interactions and their functional consequences.

Main Methods:

• Developed SSIPe, a method integrating protein interface profiles (from structural and sequence homology searches) with a physics-based energy function.
• Incorporated amino acid-specific pseudocounts to enhance profile accuracy and address database limitations.
• Evaluated SSIPe on a large dataset of 2204 mutations from 177 proteins, ensuring stringent separation of training and testing data (sequence identity < 30%).

Main Results:

• SSIPe achieved a Pearson correlation coefficient of 0.61 and a root-mean-square-error of 1.93 kcal/mol for ΔΔGbind prediction, significantly outperforming other methods.
• The combination of a physics-based energy function and knowledge-based interface profiles is key to SSIPe's superior performance.
• SSIPe demonstrated significant improvement over previous profile-based methods like BindProfX due to enhanced sequence profiles and optimized pseudocounts.

Conclusions:

• SSIPe provides a highly accurate and robust method for estimating mutation-induced binding affinity changes (ΔΔGbind).
• The developed approach enhances the prediction of mutation effects on protein-protein interactions, aiding in disease diagnosis and functional annotation.
• The SSIPe web server and source code are publicly available for broader scientific use.