[3dem] Importing Particles from Cryosparc into Relion for Processing

[Daniel Asarnow]

Hi Lauren,
Thanks for using my software. What messages do you see from Relion? You
must check both the run.out and run.err files for the extraction job - a
major hint is any "particle outside micrograph" warning.

In all our image processing suites but one (nextPYP), particle extraction
(using coordinates to duplicate the data around each particle into small
images of uniform size) is a required preprocessing step for
single-particle analysis jobs. It is up to you whether to import your new
micrographs from Relion MotionCorr back into cryoSPARC, or to extract in
relion. However, you will notice that the micrographs from Relion are
upside-down relative to the ones from cryoSPARC, because Relion flips any
.tif file when it is read into memory. Further, Relion and cryoSPARC also
use opposite conventions for the Y origin. Further further again, cryoSPARC
Import Particle jobs assume the Relion convention in the .star file - the
coordinates are inverted. Thus csparc2star.py has several options to
account for all use cases.

default : exported coordinates are inverted, so that the same micrographs
can be used for extraction in Relion (as mentioned this was the only use
case in the first versions of cryoSPARC)
--inverty : exported coordinates are not inverted, so that if they are
imported back into cryoSPARC, they will work, e.g. if you only run Class3D
in Relion
--flipy : no effect on coordinates, but flips CTF angles to account for the
upside-down images (one can always run CTF estimation again instead)

I recommend you always inspect the pick locations on the micrographs either
manually, with relion_display, or back in cryoSPARC to avoid wasting time
by extracting with wrong coordinates. They can also be verified by running
2D classification again, but it's less efficient.

The same issue will also affect the use of an empirical gain reference
produced by relion_estimate_gain. If your data is from a Gatan camera, you
will need to use the opposite gain flip (in Y) vs. the gain saved from DM
and converted by e2proc2d.py (EMAN2) or dm2mrc (IMOD).

More details, discussion, examples, etc. can be found on the pyem issue
tracker and wiki on github.
https://urldefense.com/v3/__https://github.com/asarnow/pyem__;!!Mih3wA!EYimMhBqGpR-E8kdz4pCHVhY8OP3Wtu9MEGXvKoqvtg0EkWO-hj_3sjrEvrsXcSDRyJeEfdF_W7XlDbyP5k$ 

Best,
-da

On Thu, Oct 16, 2025 at 10:55 AM Lauren Pellows via 3dem <
3dem at ncmir.ucsd.edu> wrote:

> Hi all,
>
> I am new to cryoEM field and have been working in Cryosparc for all my
> processing. I am now wanting to use Relion and have been trying to figure
> out how to import the particles I extracted in cryosparc into relion. I
> wanted to reprocess my raw micrographs in relion so I imported those ran
> motion correction and CTF estimation on those. After that I ran csparc2spar
> on my cryosparc data and now have my particles in a .star file. I also have
> gone into the particle.star file made sure the _rlnMicrographName in the
> star file to match the relion processes images. I have now imported the
> particles.star file into relion but when I run the extraction job I don't
> get any extracted particles.
>
> I have been reading online and can't find anything conclusive on if I need
> to run the extraction job. I have read that I dont need run the extraction
> job since the particles.star file should have all the metadata on the
> particles, and I can just use the particles.star to run my jobs. I also
> read that I can rerun the extraction job using the particle coordinates
> from cryosparc. Since these feel counter intuitive and I keep getting
> errors I am not sure what to do next. I know that this is a very typically
> workflow that people do so im hoping this is a pretty straightforward
> question. Any help or advice anyone has would be a great help!!
>
> Thank you so much,
>
> Lauren Pellows, PhD
>
> Postdoctoral Fellow
> Dukovic Lab
> University of Colorado, Boulder
>
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20251016/edcc60db/attachment.html>