Contextual Conformational Variability in CryoEM and CryoET using Deep Learning

Steven Ludtke
Baylor College of Medicine

By studying molecules in solution, free to explore their entropic landscapes, CryoEM takes the first step towards relaxing the rigid restraints imposed by traditional X-ray crystallography. Cellular CryoET takes the final step, by examining molecules in their true native environment, interacting with other molecules, and in their true locations and orientations within the cell. However, the most common strategies in CryoEM and CryoET analysis still focus on a single high resolution structure, or small set of discrete structures, often discarding a majority of the collected data in the process. We have developed deep learning methodologies both for exploring the population dynamics of CryoEM and CryoET data as well as additional tools for annotating cellular data to provide contextual information for each individual macromolecule. I will discuss both the deep learning strategies we use to achieve these aims as well as specific biological systems where compositional and conformational variability play a critical role in understanding function.

Presentation (PDF File)

Back to Workshop III: Cryo-Electron Microscopy and Beyond