Femtosecond protein nanocrystallography—data analysis methods

Abstract

X-ray diffraction patterns may be obtained from individual submicron protein nanocrystals using a femtosecond pulse from a free-electron X-ray laser. Many “single-shot” patterns are read out every second from a stream of nanocrystals lying in random orientations. The short pulse terminates before significant atomic (or electronic) motion commences, minimizing radiation damage. Simulated patterns for Photosystem I nanocrystals are used to develop a method for recovering structure factors from tens of thousands of snapshot patterns from nanocrystals varying in size, shape and orientation. We determine the number of shots needed for a required accuracy in structure factor measurement and resolution, and investigate the convergence of our Monte-Carlo integration method.

Coherent convergent-beam time-resolved X-ray diffraction

“…For d p ) W, the essentially plane-wave illumination shows a set of lateral shape transforms laid down on each reciprocal lattice site, in accordance with the expression for nanocrystal XRD given by Kirian et al [9]. We note that when contours of equal intensity are plotted on a plane containing the optic axis, the distribution of light for a beam at Gaussian focus is well collimated near the axis, so that a focused coherent beam need not be much wider than a nanocrystal to see the shape-transform effects described here.…”

CrystFEL: a software suite for snapshot serial crystallography

“…A fast nanocrystal diffraction simulation program, named pattern_sim, is included in CrystFEL, which simulates patterns in a manner similar to that described in the previous literature (Kirian et al, 2010). The unit-cell dimensions are taken from a Protein Data Bank (PDB; Berman et al, 2000) file; however, the structure factors themselves must be calculated by a separate means, such as using the sfall tool in CCP4 (Winn et al, 2011).…”

“…Suitable expressions for the scattering vector q are given by Kirian et al (2010). The program calculates the Laue functions in advance for the three required values of n a , n b and n c in terms of q Á a, q Á b and q Á c, respectively.…”

“…Merging of individual intensity measurements is performed by the method of Kirian et al (2010), in which the mean of all measurements for each individual reflection is taken. When the number of measurements is large, this procedure constitutes a Monte Carlo integration over the three-dimensional reflection profiles, resulting in values that are proportional to the integrated intensity.…”

“…In addition, the size and form of the shape transform vary from crystal to crystal. The Monte Carlo method of integration (Kirian et al, 2010) provides a theoretical method to process data in this situation but relies on a very high redundancy in the data set. This fact makes it necessary to index and measure intensities from many thousands of patterns, so unsupervised automatic processing of such data is of the utmost importance.…”