Post-refinement method for snapshot serial crystallography

Abstract

A post-refinement procedure has been devised for 'snapshot' diffraction data consisting entirely of partially recorded reflections, each diffraction pattern from a crystal in an orientation unrelated to the others. Initial estimates of the diffraction geometry are used to calculate initial partialities, which are then used to scale the entire dataset together to produce initial estimates of the fully integrated intensities. The geometrical parameters for each pattern are then refined to maximize the agreement between these estimates and the calculated intensities in each pattern, and the procedure repeated iteratively. The performance of the procedure was investigated using simulated data and found to yield a significant improvement in the data quality.

Keywords: data analysis; post-refinement; scaling; serial crystallography.

Figure 1.

(a) Cross section of reciprocal space showing the volume between the limiting Ewald spheres within which reflections can be excited (shaded in grey). The enlarged regions show five reflections, labelled A–E, which have different values of partiality and Lorentz factor. (b) Further enlargement of reflection ‘A’ showing the definition of r_high and r_low.

Figure 2.

Mean and maximum partialities and number of reflections in the test dataset.

Figure 3.

Comparison of R-factors with a maximum initial reciprocal space error of 0.1%.

Figure 4.

Correlation between observed and calculated partialities after (a) zero, (b) one and (c) three cycles of post-refinement.

Figure 5.

Comparison of R-factors with a maximum initial reciprocal space error of 1.0%.