GTC 2020 S21348
Presenters: Mohammad ElGamacy,Friedrich Miescher Laboratory of the Max Planck Society
Abstract
Protein drugs have revolutionized modern cancer therapeutics, as protein-based binders can effectively target molecules that are otherwise undruggable by small molecules. Presently, such binders are based on monoclonal antibodies, which are developed and engineered through lengthy and resource-intensive processes of iterative optimization that are empirically guided. Conversely, de novo protein design can offer a rational means for generating new binders with bespoke scaffolds, guided by physics-based computations. Here, we’ll present our work on developing a purpose-built, GPU-accelerated computational pipeline for designing protein-based binders de novo. As a proof of principle, we designed proteins that target a key modulator of cancer metastasis. Our experimental characterization of only a few design candidates resulted in binders with strong binding affinities. Solving the structure of one design showed atomic-level agreement between the design model and the determined structure.
Watch this session
Join in the conversation below.