Inference of germinal center evolutionary dynamics via simulation-based deep learning

Area of Science:

• Immunology
• Evolutionary Biology
• Computational Biology

Background:

• B cells and their antibodies are crucial for health and survival.
• Germinal centers (GCs) are sites of B cell maturation, involving mutation and evolution.
• A positive correlation between B cell affinity and reproduction is known but not precisely defined.

Purpose of the Study:

• To determine the exact form of the 'affinity-fitness response function' in B cells.
• To understand the relationship between antibody affinity and B cell proliferation within germinal centers.

Main Methods:

• Utilized deep learning techniques for data analysis.
• Employed simulation-based inference to model the affinity-fitness relationship.
• Conducted unique experiments replaying specific germinal center conditions in mice.

Main Results:

• Successfully learned the affinity-fitness response function.
• Quantified the relationship between B cell antigen affinity and reproductive success.
• Provided a computational model for germinal center dynamics.

Conclusions:

• The study precisely defines the affinity-fitness response function for B cells.
• This work enhances understanding of immune system evolutionary processes.
• The developed methods and open-source code facilitate further research in B cell biology.