Bug #2320
closed
Mask radius for refinements confuse pixels and angstroms
Added by Amber Herold about 11 years ago.
Updated almost 10 years ago.
Affected Version:
Appion/Leginon 2.2.0
Description
From beginning to end when running and uploading a refinement, we switch back and forth from pixels and angstroms which has caused the incorrect values to be used.
Moving forward,
- All web GUI display of mask sizes will be in angstroms.
- When launching a refinement on the python side, it takes angstroms.
- When uploading a refinement, we save the size as pixels relative to the original stack (rather than the stack created for refinement).
- The Frealign refinement takes the mask in angstroms.
- The xmipp and Eman refines take pixels.
r17495 makes changes to web gui and python launch file.
Still need to make changes to Upload to ensure mask sizes are in terms of pixels relative to the original stack box size, in uploadFrealignRefine.py, etc.
May need to pass the parent_apix and parent_boxsize from the web GUI, but better to keep it on python side.
Hi, Amber,
I thought we found out that python code takes Angstrom. The help string was incorrect, since the parameters in apRefineJobXmipp were converting all radius from Angstrom to pixel, and apRefineJobFrealign did not do any conversion.
That is correct, I miss typed it here. Too much confusion! I changed the help text to angstrom instead of pixel in the last revision, and will update the original issue text to match.
r17500 should fix xmipp and frealign upload. I have not been able to fully test it yet. May also need to do multi model refine and eman.
Found apRefineJobEMAN not converting from Angstrom input of the python script to pixel in the eman script input. r17755 fixed this.
Checked the upload and import mask of eman runs without binning in preparation. It looked good. Still need to check the binned one.
- Status changed from Assigned to In Code Review
- Status changed from In Code Review to Closed
- Affected Version changed from Appion/Leginon 2.1.0 to Appion/Leginon 2.2.0
Also available in: Atom
PDF