Deep residual networks for crystallography trained on synthetic data

Mendez, D.; Holton, J.M.; Lyubimov, A.Y.; Hollatz, S.; Mathews, I.I.; Cichosz, A.; Martirosyan, V.; Zeng, T.; Stofer, R.; Liu, R.; Song, J.; McPhillips, S.; Soltis, M.; Cohen, A.E.

doi:10.1107/S2059798323010586

Acta Crystallographica Section D
Acta Crystallographica
Section D STRUCTURAL BIOLOGY

IUCr IT WDC

home archive editors for authors for readers submit subscribe open access

view article

Figure 2
Simulated PILATUS 6M images with varying resolutions. Some parameters underlying each image are summarized in Table 1

. For each sub-image, the beam center is in the upper left corner and the resolution (determined from the B factor) is indicated by a red dashed line. The sub-images represent one quadrant of a PILATUS 6M camera, downsampled to a 512 × 512 pixel array according to Section 2.2

(see also Fig. 4

). Note that resolution here is related to the B factor by the relationship shown in Fig. 1

. Hence, while the resolution sometimes appears intuitively as the point where the scattering drops off (for example in A, B, C, D, G, H, I, K and O), at other times the Bragg reflections extend to wider angles beyond the indicated resolution (E, J, L and N). This results from inaccuracy in the resolution-versus-B factor relationship (Fig. 1

). Rarely, resolution is obscured by large background (F) or is covered by the beam stop (M). These edge cases add noise to the model training. Color bars are shown in square-root-photon units.

STRUCTURAL
BIOLOGY

ISSN: 2059-7983

Volume 80| Part 1| January 2024| Pages 26-43

https://doi.org/10.1107/S2059798323010586

Open

access

Follow Acta Cryst. D

Search IUCr Journals		doi		Advanced search
Author		volume	page