This paper develops computational methods to predict epitope overlaps in antibodies, showing improved accuracy using contrastive learning and demonstrating their utility in antibody discovery.

Patterns · Nov 13, 2025Peer-reviewedImmunology

Generation of antigen-specific paired chain antibody sequences using large language models

Perry T. Wasdin, Nicole V. Johnson, Alexis K. Janke, Sofia Held, Toma M. Marinov, Gwen Jordaan, Léna Vandenabeele, Fani Pantouli, Rebecca A. Gillespie, Matthew J. Vukovich, Clinton Holt, Jeongryeol Kim, Grant S. Hansman, Jennifer K. Logue, Helen Y. Chu, , Ivelin S. Georgiev

This paper introduces MAGE, a model that generates paired human antibody sequences against specific targets, validated against SARS-CoV-2, H5N1, and RSV-A.

bioRxiv · Dec 20, 2024 · PreprintBioinformatics

Discovery and Characterization of a Pan-betacoronavirus S2-binding antibody

Nicole V. Johnson, Steven C. Wall, Kevin J. Kramer, Clinton Holt, Sivakumar Periasamy, Simone I. Richardson, Nelia P. Manamela, Naveenchandra Suryadevara, Emanuele Andreano, Ida Paciello, Giulio Pierleoni, Giulia Piccini, Ying Huang, Pan Ge, James D. Allen, , Ivelin S. Georgiev

The paper identifies a pan-betacoronavirus antibody that binds the S2 subunit of spike proteins and induces Fc-dependent antiviral responses, offering insights into non-neutralizing antibody protection.

Structure · Sep 25, 2024Peer-reviewedMolecular Biology