Warp-PyTom-Relion5 integration

[Kareem Farghaly]

Dear Warp community,

I'm working on a subtomogram averaging workflow and encountering an issue during the integration of Warp, PyTom, and Relion5. Below is a description of my current pipeline and the problem I’m facing.

Tomogram Reconstruction:
I reconstructed a single tomogram using Warp.

Template Matching (PyTom):
I used PyTom to perform template matching for ribosomes with the following command:
pytom_extract_candidates.py -j results/MGS001_T1_ts_001_13.51Apx_job.json --cut-off 0.15 -n 1000
**I have also used the --relion5-compat version but the end result is the same.

Exporting Particles (WarpTools):
I attempted to export the matched particles using WarpTools ts_export_particles:

WarpTools ts_export_particles \
  --settings warp_tiltseries.settings \
  --input_star PyTom/results/MGS001_T1_ts_001_13.51Apx_particles.star \
  --output_star relion/matching.star \
  --output_angpix 13.508 \
  --coords_angpix 13.508 \
  --box 64 \
  --diameter 300 \
  --relative_output_paths \
  --2d

Initially, this failed until I set the _rlnTomoName values to match the name of the corresponding .tomostar file from Warp in the .star file.

After making these changes, the ts_export_particles command ran successfully and generated matching.star and matching_tomograms.star.

Subtomogram Extraction:
In Relion5, I used an Extract Subtomos job with: matching.star as the input particle set and matching_tomograms.star as the input tomogram set

And this is a photo of the Reconstruct tab.

When I run the extraction job in Relion5, I get this and the extraction doesn’t work. I have also tried different box sizes

The particles are removed because they were too close to the edges.

The unbinned dimensions of the tomogram are 11520x8184x3000 with a pixel size of 1.689.

How can I solve this problem and is the way I am trying to integrate Warp and Relion correct or is there a better way?

Any help or insights would be greatly appreciated! Thank you.

Best regards,

Kareem

[Juha Huiskonen]

Hi Kareem, 

You don't need to extract the particles as Warp has already extracted them.  Just proceed directly to 3D classification or 3D auto-refine. Use the STAR files generated by Warp is inputs (you need to browse up to the RELION main folder to see them).

Best regards,

Juha