Abstract:

One of the holy grails of drug-design is to obtain high-selectivity binding of a drug to the target protein, an extremely challenging endeavor for proteins that share high sequence and structural similarity. We investigate several systems of high biological importance to elucidate how selectivity emerged during evolution and how it can inform the design of selective drugs. In particular, we investigate three protein families, i.e. human protein kinases, receptors for plant hormone, and strigolactone and SWEET transporters. These proteins show several orders of magnitude difference in their ligand affinity, but share high sequence and structural similarity. This task involves understanding the structure-function relationship of proteins during evolution using advanced computational methods. We show how evolution-based methods such as the combination of reinforcement learning-based adaptive sampling and evolutionary couplings developed in our group could help simulate rare conformational change processes in proteins. In this talk, we will not only demonstrate the utility of these new techniques to address a key challenge of understanding evolution of ligand perception, but also the need, feasibility and impact of computational evolutionary biophysics in understanding the fundamental questions in biology.