mCSM-PPI2

mCSM-PPI2 predicts the effects of missense mutations on protein-protein interaction (PPI) binding affinity using graph-based structural signatures and machine-learning models.

Key Features:

• Graph-Based Structural Signatures: Employs graph-based structural signatures to model the effects of variations on inter-residue interaction networks.
• Integration of Evolutionary and Network Data: Incorporates evolutionary information and complex network metrics into the predictive framework.
• Energetic Terms and Optimized Predictor: Integrates energetic terms with other features to produce an optimized predictor that has demonstrated superior performance in comparative evaluations such as CAPRI blind tests.

Scientific Applications:

• Disease variant interpretation: Identify potential disease-causing missense mutations by assessing their impact on PPIs.
• Therapeutic design: Aid design of therapeutic interventions targeting specific protein interactions by predicting mutation-induced affinity changes.
• Drug discovery: Facilitate drug discovery by predicting how mutations may affect target binding sites and interface stability.

Methodology:

Encodes mutations with graph-based structural signatures and integrates evolutionary information, complex network metrics and energetic terms within machine-learning predictive models.