Molecular dynamics flexible fitting is popularly employed as a structure-refinement tool. However it has clear limitations for ensemble refinement and often spurious geometries are determined stemming from the underlying force fields. A new MD-based protector ensemble refinement tool is presented called CryoFold. In addition to the most probable interpretation of the data, CryoFold generates rare conformations that are often required to address the regions of poor local resolution, such as ligand binding pockets or solvated interfaces . It avoids over-fitting artifacts, which confirms accurate geometries. Most importantly, weights of different members of a cryo-EM data-guided ensemble offers a description of the conformational landscape. A highly parallel implementation of the workflow is available for supercomputing applications that now extends molecular dynamics refinements to cryo-ET datasets.
Copyright. All Rights Reserved.