Daniel Faissol

This paper introduces an integer linear programming method for optimizing antibody libraries by balancing diversity and affinity, showing improved performance over baseline designs.

AbBERT is a transformer-based language model trained on antibody sequences to predict humanness, showing utility in antibody design and optimization.

The paper predicts the SARS-CoV-2 Omicron variant has higher ACE-2 binding affinity and reduced neutralizing antibody binding compared to Delta and other variants using computational methods.

This paper uses machine learning and supercomputing to design antibodies targeting SARS-CoV-2, generating 20 antibody sequences with improved binding energy to the RBD.