Significantly enhancing human antibody affinity via deep learning and computational biology-guided single-point...

Area of Science:

• Computational biology
• Immunology
• Drug discovery

Background:

• Antibody affinity enhancement is crucial for therapeutic efficacy and safety.
• Traditional mutation methods struggle with vast sequence spaces.
• Current limitations necessitate advanced computational approaches for antibody engineering.

Purpose of the Study:

• To develop a novel computational pipeline for identifying affinity-enhancing antibody mutations.
• To leverage deep learning and evolutionary constraints for precise mutation prediction.
• To demonstrate the pipeline's generalizability across different antibody targets.

Main Methods:

• Integrated computational pipeline: evolutionary constraints, statistical potentials, molecular dynamics, metadynamics.
• Deep learning framework: MicroMutate for microenvironment-specific mutations, graph-based models for binding probabilities.
• Screening of single-point mutant antibodies against H7N9 influenza hemagglutinin and death receptor 5.

Main Results:

• Identified a mutant antibody targeting H7N9 hemagglutinin with a 4.62-fold affinity improvement.
• Engineered an antibody against death receptor 5 with a 2.07-fold increase in affinity.
• Demonstrated successful application in identifying high-affinity mutants from subnanomolar antibodies.

Conclusions:

• The integrated computational pipeline effectively identifies affinity-enhancing mutations.
• Deep learning and computational methods offer a powerful platform for antibody design.
• This approach accelerates the discovery of improved antibody therapeutics while maintaining immunogenicity and expression.