Saved Particle Lists are written in angstroms and not pixels

[dmichalak]

Importing a Particle List created in 3dmod (.mod) is displayed correctly on the tomogram file loaded into ArtiaX. However, when saving the Particle List as a new .mod file, the coordinates are scaled by the Angstrom/pixel value. This was confirmed by comparing the coordinates written by IMOD's model2point on the .mod file generated by 3dmod and the saved .mod file from ArtiaX:

$ model2point 3dmod.mod model2point_3dmod.coords
$ head -n 1 model2point_3dmod.coords
551.33 966.00 317.00

$ model2point ArtiaX.mod model2point_ArtiaX.coords
$ head -n 1 model2point_chmx.coords
4774.55 8365.56 2745.22

For this data, there is 8.66 Angstroms/pixel.

[uermel]

Hi, thanks for the report. It seems I didn't run into a model file with a pixel size specified while testing this (I used rather old files from the PEET example datasets).

I looked into the model format again, and there is both a scaling factor for x, y and z and a pixel size defined. I've committed a solution now in the development branch, which reads the pixel size and scaling factors and computes a final (isotropic) pixel size in angstrom. This is then set as the scale factor for particle origins in the "Select/Manipulate" tab.

Saved coordinates will be in pixels, and the pixelsize will be set to x_scale*pixelsize with units of Angstrom. Is this an acceptable solution?

You can try it using the attached build (extract the wheel file before installing). I'm still working on finding a way to set up daily builds.

ChimeraX_ArtiaX-0.2a1-py3-none-any.zip

[dmichalak]

Yes, that sounds good! Apologies for the daily in response. I apparently didn't set up notifications correctly to see when there was a reply to this.