Bug #1225
closedViper to EMAN conversion option crashes upload model job
0%
Manually run two EMAN commands:
proc3d viperfile.mrc tmpfile.mrc icos5fTo2f
then
proc3d tmpfile.mrc emanfile.mrc rot=0,0,90
These should be accessible as a tool tip or link next to the option box. This information was difficult to find.
Description
Hi Jeff,
The Viper to Eman conversion error must be a bug. I had never tried it before today, and I got the same error as you. This needs to be looked into, so you can submit a redmine issue for it if you want.
Lauren
On Mar 4, 2011, at 3:27 PM, Jeff Speir wrote:
Hi Lauren,
I thought of this and I've tried twice (once yesterday, once today) to upload a model in viper orientation and rotate it to EMAN orientation using the checkbox in the Upload Model page. Unfortunately it has crashed both times. Please see:
ami/data17/appion/11jan12a/models/accepted/file_11mar02p48
/ami/data17/appion/11jan12a/models/accepted/file_11mar03k45
Looks like the same error both times, and may be related to using the viper to eman option (?). The model I uploaded previously is identical except already rotated - I used the same parameters to upload but had not checked the rotation option.
Updated by Lauren Fisher over 13 years ago
On Mar 14, 2011, at 2:29 PM, Jeff Speir wrote:
Hi Lauren,
I had sent you the commands Gabe found for converting virus maps in viper icos orientation to EMAN icos orientation. Turns out there was a slight error. The 90 degree rotation in the second command should be -90. Didn't unravel this until EMAN spit back a good enough reconstruction to see that the icos symmetry wasn't exactly matching the model that went in. There also appears to be a bug in EMAN pdb2mrc such that when you specify low resolution for the output map it slows way, way down (final map is fine, just takes far too long to compute). A workaround was to produce the default map (1Å spacing) then low pass filter it. The simple script below embodies both fixes and is what I used. Thought it might be useful for you to know how I worked from Viper coordinates to an EMAN starting model in the end.
Cheers,
Jeff
#!/bin/tcsh
- Compute map from model in Viper std icos orientation and
- rotate to EMAN std icos orientation
- generate map, 1A/voxel
pdb2mrc afhv_vprstd_particle.pdb tmp1.mrc
- low pass filter map, lp=A/voxel (i.e. resolution)
proc3d tmp1.mrc tmp2.mrc apix=1.0 lp=25 edgenorm
- rotate map from Viper std to EMAN std icos orientation
proc3d tmp2.mrc tmp3.mrc icos5fTo2f
proc3d tmp3.mrc afhv_25A_eman.mrc rot=0,0,-90
- clean up intermediate maps
/bin/rm tmp1.mrc tmp2.mrc tmp3.mrc
Updated by Lauren Fisher over 13 years ago
Error message from /ami/data17/appion/11jan12a/models/accepted/file_11mar02p48/file_11mar02p48.appionsub.log:
... Time stamp: 11mar02p49 ... Function name: uploadModel ... Appion directory: /usr/lib64/python2.4/site-packages !!! WARNING: Using split database Connected to database: 'ap292' ... Committing data to database ... Found project id=292 for session 11jan12a in 6.81 msec ... Looking up symmetry for input: 2 ... Selected symmetry group: ICOS -- Icos (5 3 2) EMAN ... Run directory: /ami/data17/appion/11jan12a/models/accepted/file_11mar02p48 !!! WARNING: directory '/ami/data17/appion/11jan12a/models/accepted/file_11mar02p48' already exists. ... Writing function log to: uploadModel.log ... Found 4 processors on this machine ... Running Appion version 'r15281' Naming density model: /ami/data17/appion/11jan12a/models/accepted/file_11mar02p48/upload-vni206_model_25A_vpr-11mar02p49 ... copying original model to a new location: /ami/data17/appion/11jan12a/models/accepted/file_11mar02p48/upload-vni206_model_25A_vpr-11mar02p49.mrc calculating dimensions... Traceback (most recent call last): File "/opt/appion/bin/uploadModel.py", line 234, in ? uploadModel.start() File "/opt/appion/bin/uploadModel.py", line 217, in start apVolume.viper2eman(mrcname, mrcname, apix=self.params['apix']) KeyError: 'apix'
Updated by Lauren Fisher over 13 years ago
- Status changed from New to In Code Review
- Assignee set to Arne Moeller
The variable apix was changed to newapix at some point and the viper2eman command was not updated to reflect this change. r15520 fixes this bug.
For future reference the EMAN commands to convert Viper orientation to Eman orientation are:
proc3d input output rot=0,0,90 ###rotates map 90 degrees about z-axis
proc3d input output icos2fTo5f ###moves from 2fold facing z to 5fold facing z
proc3d input output origin=0,0,0, apix=1.64 ###resets origin and pixel size
Updated by Anchi Cheng almost 9 years ago
- Status changed from In Code Review to Closed