CryoDRGN-ET subtomogram analysis
Heterogeneous reconstruction of cryo-ET subtomograms
Now available in production/stable release in cryodrgn version 3.3.0! See News and release notes
CryoDRGN-ET is available in version 3.0.0+ through the same cryodrgn software interface with additional flags for subtomogram analysis. See our preprint here for a description of cryoDRGN-ET and associated results.
Preprocessing
Currently, cryoDRGN-ET expects 2D particle tilt series images in a .star
file that is exported from Warp/M or RELION5.
Downsample images
Parse pose information
Parse CTF information
Test inputs are correctly parsed with traditional homogeneous reconstruction (GPU required)
CryoDRGN-ET training
Example command:
In particular:
--encode-mode tilt
is required to properly treat tilt series data--dose-per-tilt
and--angle-per-tilt
are used for dose exposure correction. The default value of--angle-per-tilt
is 3 degrees and is left off of the example command.In our current experiments, we use a KL regularization weight of
--beta 0.025
Training a model on 16,655 particles for 50 epochs on 1 A100 GPU took 3h, 38min.
Analysis
Once a cryoDRGN-ET model has finished training, use cryodrgn analyze
to visualize the latent space and generate volumes.
This portion of the analysis is similar to the workflow in single particle cryodrgn. See the EMPIAR-10076 tutorial for further documentation.
Feedback
Please file a github issue or contact Ellen (zhonge@princeton.edu) with any questions or feedback!
Last updated