Electron cryomicroscopy (cryoEM) is a maturing field in structural biology that can determine structures of macromolecular machines and cells at a broad range of resolution from 4 to 100 Å. Evidence to the growth of the field, the increasing number of publications in cryoEM has prompted interest from the PDB and EBI to archive density maps and associated models derived from cryoEM (EMDB). Generally, the molecular mass of the biological machine is on the order of 1-100 MDa, which is often too difficult to study by conventional X-ray crystallography. CryoEM is an ideal technique to bridge the information gap between cell biology and crystallography/NMR of individual molecular components of biological machines including viruses, chaperonins, ion channels, ribosomes, transporters, enzymes, filaments and bundles. Single-particle cryoEM can be used to explore the complex and dynamic behavior of individual biological machines in different functional states. However, the resulting data from cryoEM experiments are presently limited to medium (5-10Å) to low (10-20Å) resolutions. This limited resolution is due to several factors including sample heterogeneity due to conformational flexibility.

Here, we hypothesize that single particle cryoEM images contain data of mixed conformations, which can be sorted out computationally to derive multiple models from different subsets of particle images. Therefore, the theme of this proposal is to develop a physics-based computational methodology using SimTK tools to derive an ensemble of potential structural models, which we hypothesize would represent the dynamic nature of the biological machine itself.