Java-Based Platform For Immuno-Engineering Biotherapeutics


At a global pharmaceutical company, Gordon led a project to develop a combined laboratory and computational platform for predicting the immune response of patients to therapeutic proteins and antibodies administered in the clinic, and for re-engineering these molecules to hide them from the immune system. The prototype was developed in Python and the modeled immune response was calibrated using laboratory assays.

The production version of the platform was re-implemented in Java and eventually deployed throughout the global company, becoming a standard part of the drug development workflow for all new biological drug projects. Once the software had identified the potentially immunogenic regions of a molecule, it then drew upon a variety of sources of information, including BLAST alignments, models of evolutionary conservation, and three-dimensional structural information, to guide the protein engineer in the process of replacing these immunogenic regions, while most importantly, also preserving the structure and function of the protein.