Modeling Protein-Ligand Cryo-EM Structures with Neural Network Potentials

BBSRC : Regan Harle : BB/T008709/1

Drugs act by binding to a target protein. Determining the three-dimensional shape the complex between a drug and a protein is challenging, particularly when the protein spans the membrane separating the interior and exterior of a cell. Cryo-electron microscopy allows these structures to be measured experimentally, but often the resolution is too low to aid the development of new drugs. We will combine this technique with artificial intelligence protein structure determination and computer simulations to determine the structures of drug-bound structures with minimal direction by the user but with a level of resolution that had not been possible in the past. I will be trained in machine learning, molecular simulation, and artificial-intelligence protein structure prediction.