view article
Figure 1
(a) Flowchart of the orientation determination algorithm as well as the 3D phase recovery algorithm. The orientation determination algorithm begins with a roughly accurate predicted model, which provides an initial reference Fourier volume. The volume will be iteratively improved with updated orientation information. The above procedure will be performed multiple times until convergence is reached. The predicted model also provides an initial set of phases instead of random ones for the conventional phase retrieval algorithm, which iterates back and forth in dual-space to recover the final electron-density map. (b) Definition of rotation, with three Euler angles α, β and γ under a kinematic coordinate system. More specifically, the first angle α indicates rotation around the blue z axis, which also brings the blue x axis to the green N axis. The second angle β indicates rotation around the green N axis, which further brings the blue z axis to the red Z axis. The third angle γ indicates the rotation around the red Z axis. If we imagine that the circle represents a 2D Fourier slice before and after 3D rotation, then the Z axis becomes its normal axis. Of note, the incident X-ray is in the red Z axis direction. (c) Example of the simulated diffraction pattern in this study, and the upper and lower limits of the resolution circles for calculating Pearson correlation are also shown.

IUCrJ
Volume 11| Part 4| July 2024| Pages 602-619
ISSN: 2052-2525
https://doi.org/10.1107/S2052252524004858