Abstract
The advent of the X-ray free-electron laser (XFEL) has made it possible to record diffraction snapshots of biological entities injected into the X-ray beam before the onset of radiation damage. Algorithmic means must then be used to determine the snapshot orientations and thence the three-dimensional structure of the object. Existing Bayesian approaches are limited in reconstruction resolution typically to 1/10 of the object diameter, with the computational expense increasing as the eighth power of the ratio of diameter to resolution. We present an approach capable of exploiting object symmetries to recover three-dimensional structure to high resolution, and thus reconstruct the structure of the satellite tobacco necrosis virus to atomic level. Our approach offers the highest reconstruction resolution for XFEL snapshots to date and provides a potentially powerful alternative route for analysis of data from crystalline and nano-crystalline objects.
Highlights
Ultrashort pulses from X-ray free-electron lasers (XFELs) have recently made it possible to record snapshots before the object is damaged by the intense pulse [1,2]
The advent of the X-ray free-electron laser (XFEL) has made it possible to record diffraction snapshots of biological entities injected into the X-ray beam before the onset of radiation damage
We present an approach capable of determining structure from diffraction snapshots of symmetric objects to 1/100 of the object diameter and demonstrate threedimensional structure recovery to atomic resolution from simulated noisy snapshots of the satellite tobacco necrosis virus (STNV) at the signal level expected from viruses currently under study at the LCLS X-ray free-electron laser
Summary
Ultrashort pulses from X-ray free-electron lasers (XFELs) have recently made it possible to record snapshots before the object is damaged by the intense pulse [1,2]. This has, for example, resulted in de novo determination of protein structure from nano-crystals fabricated in vivo [3]. The ultimate goal, remains the determination of the three-dimensional structure of individual proteins and viruses [4] and their conformations [5] This requires the ability to recover structure from an ensemble of ultralow-signal diffraction snapshots of unknown orientation. More recent methods offer improved performance, either by obviating the need for iteration [13], or by improved scaling per iteration, e.g. (R5 log R) [14], though the magnitude and scaling of the number of iterations, where needed, remain unknown
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
