Protein-protein interactions play a significant role in many biological processes, from governing cellular functions to mediating signaling pathways. The design of polypeptide binders-chains of amino acids engineered to selectively bind a target protein-is an important area of research with broad scientific and therapeutic applications. However, current state-of-the-art generative models often lack the ability to design binders with high specificity. Because these models optimize binding to a single target without explicit consideration of structurally similar off-target proteins, they risk producing binders that interact with unintended targets, leading to unwanted side effects. Our project addresses this limitation by developing computational methods that optimize for selectivity, maximizing affinity for the intended target while minimizing affinity for competing off-target proteins, with the goal of improving the applicability of binders in protein engineering, therapeutics, and biological research.

I am participating in SuperUROP to build on my interests in computer science, biology, and protein design through a longer-term research project. I am also looking forward to the formalized training in writing and presenting papers, as these are skills I want to develop as a researcher.