Briney Lab

We generate massive datasets of human antibody sequences and use them to train state-of-the-art models of structure, function and evolution. Our group works at the interface between experimental and computational immunology, which ideally positions us to iterate rapidly through successive train–test–learn cycles, a process that requires seamless transitions from the bench to the GPU (and back). The models and methods we create can be used to evaluate antibody responses to candidate vaccines, engineer monoclonal antibodies to improve a variety of properties (affinity, expression, developability, etc.), and fully de novo discovery of antigen-specific antibodies directly from sequence.