Beyond integration: modeling every pixel to obtain better structure factors from stills

Area of Science:

• Crystallography
• Structural Biology
• X-ray Science

Background:

• Traditional crystallographic data processing relies on pixel integration.
• Serial femtosecond crystallography (SFX) averages many exposures, which is slow and hinders accurate anomalous difference measurement.

Purpose of the Study:

• To develop a novel approach for improving the accuracy of structure factors from SFX data.
• To enable more precise measurement of anomalous differences in SFX.

Main Methods:

• A physical model was created to describe and decouple contributions to pixel intensities in SFX.
• Maximum likelihood estimation optimized model parameters, including lattice orientation, unit-cell dimensions, and structure factor amplitudes.
• Prior knowledge of positive structure factor amplitudes was incorporated via reparameterization.

Main Results:

• The novel method accurately determines anomalous structure factors from SFX data.
• The approach requires an order-of-magnitude fewer shots compared to conventional integration methods for similar accuracy.
• Tested on a synthesized ytterbium(III) lysozyme dataset, the method showed stable anomalous difference signals.

Conclusions:

• This new physical modeling approach significantly enhances structure factor accuracy in SFX.
• The method offers a more efficient way to collect and process SFX data, especially for anomalous scattering.
• It paves the way for more accurate and faster structural determination using SFX.