Oct 16, 2023
Image processing and 3D reconstruction
- 1Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720, USA.
Protocol Citation: Minghao Chen 2023. Image processing and 3D reconstruction. protocols.io https://dx.doi.org/10.17504/protocols.io.x54v9d99mg3e/v1
License: This is an open access protocol distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Protocol status: Working
We use this protocol and it's working
Created: June 05, 2023
Last Modified: May 31, 2024
Protocol Integer ID: 82922
Keywords: ASAPCRN
Abstract
Image processing and 3D reconstruction
Image processing
Image processing
Use cryoSPARC for the following steps except those particularly mentioned.
Do motion correction by [Patch Motion Correction]
Bin 2x in fourier cropping for super-resolution video stacks
Bin 1x in fourier cropping for regular video stacks
Do contrast transfer function determination by [Patch CTF Estimation]
Remove the outlier micrographs base on the estimated defocus and resolution value.
Do particle picking by [Topaz]
Manually pick 10 micrographs as learning dataset
Optimize the 'picking threshold' with the 10 mics
Apply the parameter to the entile dataset
Particle extraction
Use the box size 1.5 times larger than the target particles
Bin 4x to facilitate the following classification jobs
2D classification
Set 50-100 classes dependent on the data size
Remove the obvious junk particles
Obtain an initial model
[1] Use Ab-initial (Optional) only select the 2D classes that show high-resolution features
[2] Use previously determined structure if it's available
[3] Create a new medel by AlphaFold
Do 3D classification by [Reterogeneous Refinement]
Low-pass your model to 15-20 Å
Run the job with 2-3 junk models
Run multiple times (typically 2-4 rounds) until the result converges
Re-extract the particles with
bin 2x for super-resolution video stacks
bin 1x for regular video stacks
3D reconstruction
3D reconstruction
Do 3D reconstruction by [Homogeneous Refinement]
Repeat 2-3 times until the resolution converges
Check whether the FSC curve is healthy
(Optional)
Do CTF refiment followed by homogeneous refinement.
Check whether the resolution get improved
(Optional)
Do local refinement if the map contains multiple rigid sub-regions
Decide the masks based on [3D Variability] or [3D Flex]
Use [ChimeraX] to create the maps
Use [EMAN2] to compose the final maps at the end