Antibody attributes that predict the neutralization and effector function of polyclonal responses to SARS-CoV-2

Area of Science:

• Immunology
• Virology
• Machine Learning

Background:

• Antibodies offer protection against SARS-CoV-2 but key immunity attributes are unclear.
• Understanding the humoral response is crucial for developing effective vaccines and therapies.

Purpose of the Study:

• To use machine learning to predict antibody effector functions and neutralization potency.
• To gain insights into the mechanisms of antibody action against SARS-CoV-2.

Main Methods:

• Machine learning models were developed to correlate polyclonal antibody response characteristics with effector functions and neutralization.
• Biophysical antibody profiles were used to predict antibody-mediated phagocytosis, cytotoxicity, complement deposition, and neutralization.
• Discovery and validation cohorts were used to ensure model accuracy.

Main Results:

• Accurate prediction models for antibody effector functions and neutralization were established.
• SARS-CoV-2-specific IgM was identified as a significant predictor of neutralization activity.
• Experimental depletion confirmed the mechanistic role of IgM in neutralization.

Conclusions:

• Validated models highlight desirable attributes of the humoral response for vaccine design.
• These findings can guide the development of improved vaccines and antibody-based interventions against SARS-CoV-2.