CSM-AB

CSM-AB predicts antibody–antigen binding affinities and provides docking scoring functions to evaluate and rank antibody–antigen complex poses for immunotherapy-related research.

Key Features:

• Graph-Based Signatures: Utilizes graph-based signatures to model the interaction interfaces of antibody–antigen complexes.
• Machine Learning Models: Employs machine learning models trained on known antibody–antigen complexes to predict binding affinities, achieving a Pearson's correlation coefficient of up to 0.64 in blind tests.
• Docking Scoring: Functions as a docking scoring tool to accurately rank near-native poses of antibody–antigen complexes.

Scientific Applications:

• Immunotherapy Development: Supports assessment of antibody–antigen binding affinities to inform therapeutic antibody selection and optimization.
• Docking Pose Evaluation: Ranks and discriminates near-native docking poses for structure-based studies of antibody–antigen complexes.
• Experimental Design Guidance: Provides affinity predictions to prioritize variants and guide experimental validation strategies.
• Antibody Engineering: Aids design and selection of antibodies with enhanced binding affinity and specificity.

Methodology:

Transforms antibody–antigen interaction interfaces into graph-based signatures that serve as inputs to machine learning models trained on known complexes to predict binding affinities and to score and rank docking poses.