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 15
Maximum composite images for the first 20 exposures (4° total rotation) from the SSRL data sets indicated by the subplot labels B, H, N, D, I and O. Based on Resonet overlapping lattice detection (see Fig. 14

) it was known that data sets B, H and N had a high chance of containing overlapping lattice scattering (80%, 74% and 57%, respectively). On the contrary, data sets D, I and O had a low chance of containing overlapping lattice scattering (12%, 0% and 0%, respectively), but were similar in setup to experiments B, H and N. All color bars are in `square-root-photon' units. While data set N clearly exhibited diffraction indicative of overlapping lattice features from separate protein crystals, it appeared that data sets B and H contain more subtle features (streaks and Bragg peaks from ice and/or salt) that cause Resonet to infer the presence of overlapping lattices.

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