Will White

Research/Areas of Interest

Antibody-based therapeutics have been shown to be incredibly effective at treating a wide range of medical conditions, from the use of antisera in neutralizing toxins, to monoclonal antibodies that treat autoimmune disease. However, classical antibodies can be difficult to manufacture because of their complex and dimeric structure. Additionally, although traditional immunization-based methods for antibody development have been successful, it can be difficult to optimize a lead candidate in a rational way because the key interactions formed by antibody loops are challenging to predict or model. My work aims to tackle these issues, first by using camelid antibodies which achieve comparable binding to classical antibodies using a simpler monomeric structure, and second by developing deep learning methods to predict camelid antibody binding to a variety of targets. Specifically, I am working with data from sequencing of camelid antibodies from alpacas immunized with antigens derived from pathogens including CoV2, botulinum, and schistosomes (a parasite) to build models that can predict binding strength from the antibody sequence. These methods should reduce the experimental burden to select promising candidates, and help to create cheaper and more efficacious antibody therapeutics.

Advisors:

Lenore Cowen, PhD
Charles Shoemaker, PhD