Speaker: Jason Kaelber, Ph.D.
Rutgers Institute for Quantitative Biomedicine
Rutgers University
Title: Three CryoEM stories: discovery of a new pathogen, escaping pseudosymmetry traps, and limitations to resolution
To crack a nationwide pandemic in farmed beetles, we used cryoEM to directly identify the previously-unknown causative virus without genome sequencing. We confirmed etiological association by in vivo experiments. The genomic ssDNA of this virus has an unusual degree of ordering and is closely-associated with the capsid interior. The diagnostic workflow provides proof-of-concept that could be applied to outbreaks in humans or other species. Most cryo-EM structures are solved by projection matching. This approach is vulnerable to getting stuck in local minima, including pseudosymmetry traps, and even a global minimum may reflect noise bias. An alternative to 3D reconstruction is to directly test structural hypotheses over the set of single particle observations. We have implemented such a test in a naturally simple case in which a small number of structural hypotheses can be enumerated with essentially identical power spectra.
Classic cryoEM information theory posits that the number of particles is proportional to eB/2d² where B is the temperature factor and d is the resolution, and B will be improved by superior microscope/detector quality and higher operating voltage. For a representative bacterial enzyme, we show that this relationship is constant up to a critical particle count but not a critical resolution. We propose that plots of this relationship should be created on-the-fly as an experiment termination criterion.
