Maximizing information in highly complex cryo-EM data

Cryo-electron microscopy single particle reconstruction (cryo-EM SPR) provides detailed structural information for macromolecules by averaging many images of individual molecules. However, accurate image alignment is a bottleneck due to the weakness of signal generated by individual molecules. Therefore, in practice, structures of only relatively large macromolecules can be determined by cryo-EM SPR. Phase plates are devices that enhance the signal in cryo-EM images, but current approaches to data collection do not allow us to capitalize on this potential enhancement. In order to achieve dramatically higher contrast, a much more precise data collection setup is required. The overall objective of this Phase I SBIR proposal is to develop enabling technologies for optimal use of the phase plate so that its awaited advantages become available. For practical reasons, this requires ultra-fast analysis of huge image datasets in real-time, necessitating the development of GPU-based algorithms. These methods will enable highly efficient data collection with the high level of precision necessary to use phase plates close-to-focus when the contrast enhancement is the highest. We will develop, implement, and test multiple calibration procedures, with the faster ones performed for each data collection area while slower, more elaborate ones are performed as needed during data collection. Together, this approach will produce a module that can be integrated into standard data collection schemes. In Phase I, we will develop a proof-of-principle implementation, while in Phase II and later we will integrate this module into complete data collection software package that will be commercially distributed. During and after Phase II, we will also validate the approach in a broad range of experimental conditions, expanding functionality as needed. The redesign of the approach to data collection using a phase plate will benefit all types of cryo-EM SPR and fiber reconstruction: (1) the resolution of all reconstructions will improve significantly, (2) cryo-EM will be applicable to projects where particles have lower mass, and (3) molecular motions for larger particles can be better identified and analyzed. These benefits are of high significance for this quickly expanding field, so the potential for successful commercialization is high.