SAAMBE-SEQ: a sequence-based method for predicting mutation effect on protein-protein binding affinity.

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• Many genetic disorders stem from mutations altering protein-protein interactions and binding affinity.
• Accurate prediction of mutation effects on binding free energy (ΔΔG) is crucial for therapeutic development.
• Existing structure-based methods are limited for genome-scale studies due to the lack of structural data for many proteins.

Purpose of the Study:

• To develop a novel, sequence-based computational method for predicting the impact of mutations on protein-protein binding free energy.
• To create a tool applicable to genome-scale investigations, overcoming the limitations of structure-dependent approaches.

Main Methods:

• Developed SAAMBE-SEQ, a Gradient Boosting Decision Tree machine learning algorithm.
• SAAMBE-SEQ utilizes 80 features, including evolutionary information, sequence properties, and physical property changes at mutation sites.
• The method requires only amino acid sequence information, not 3D structural data.

Main Results:

• Achieved a Pearson Correlation Coefficient (PCC) of 0.83 in 5-fold cross-validation against experimental ΔΔG data.
• Demonstrated strong performance on a blind test set (no-STRUC) of mutations lacking structural information, with PCCs ranging from 0.37-0.46.
• SAAMBE-SEQ's accuracy is comparable or superior to advanced structure-based methods and offers significant speed advantages.

Conclusions:

• SAAMBE-SEQ provides an accurate and efficient sequence-based approach for assessing mutation effects on protein-protein interactions.
• Its applicability to genome-scale analyses makes it a valuable tool for understanding genetic disorders and developing therapies.
• The method is readily accessible via a webserver and stand-alone code.