cs2star 0.7.0

Description:

cs2star 0.7.0

cs2star





A simple utility that wraps around csparc2star.py to convert particle positions from Cryosparc JOB directories into a RELION-ready directory. On top of what csparc2star.py already does, this script will traverse the Cryosparc job tree to find the relevant .cs files, automatically symlinks (or copies) the .mrc files (renaming them as appropriate to mrcs and updating the rlnMicrographName column to reflect the change). cs2star also separates micrograph information in a micrographs.star file, which is required in several relion jobs.
Installation
This package requires pyem, which is not available on pip under that name (there is a pyem on pip; that's a different package!). To install it manually [NOTE: some fixes necessary for this are not yet on the main repo, so we're installing from a fork]:
pip install git+https://github.com/brisvag/pyem.git

You can then install cs2star with pip:
pip install cs2star

Usage
cs2star cryosparc_job_dir [dest_dir]

A couple more options are available, check out the help page with -h.
Example
Assuming Cryosparc/J23 is a classification job:
cs2star Cryosparc/J23 -p

will create in the working directory:

a relion-ready particles.star
a relion-ready micrographs.star
a patches directory containing the particle images to be used by relion refinement
You can now run relion and provide particles.star and micrographs.star as the inputs.

Tips

While you can use the --class option to select specific classes from a cryosparc job, the class indexes do not correspond the the class order in Cryosparc (which is simply based on particle number). A much easier way to select specific classes to convert to .star is to use the select 2D classes job in cryosparc, then provide this job as input to cs2star.
--swapxy defaults to True. This is because usually that's what you need; unfortunately, this may differ on a case by case basis, so make sure to check the data after conversion with you favourite viewer to ensure that particles are correctly centered.