Bug #2406
openConvention of Ellipse angles and CTF astigmatism angles
0%
Description
create new bug, so slow migration of our incorrect CTF astigmatism angle convention, this should probably wait until the end of the CTF Estimation Challenge.
Updated by Anchi Cheng over 11 years ago
- Assignee set to Neil Voss
- Affected Version changed from Appion/Leginon 2.1.0 to Appion/Leginon 3.0.0
Neil,
I am going to work on this this week. First thing I found is your change in r17701. The comment in line 82 to line 84 defines a as the major axis of the ellipse, alpha is measure from x-axis to a (major axis. Then in r17701, you reverse axis when a > b, which makes alpha measured from x-axis to minor axis.
What is it really?
Updated by Neil Voss over 11 years ago
Yeah, I think the ellipse angle (which is based on a list of x and y values) is opposite my CTF angle (which is based on image data).
Updated by Anchi Cheng over 11 years ago
Neil,
Your r17800 broke Appion. Until the astigmation angle input to that is changed, the sign change in the display should not be done, yet. I reverted it back in r17804, but will use your change to get the conversion ready. What is coded in r17701 worked with the old code of ctfdisplay. I was just asking a question, mainly wonder if you want to change the comment in the code.
Updated by Anchi Cheng about 11 years ago
- Assignee changed from Neil Voss to Anchi Cheng
I've made the sign changes in my sandbox and created a separate aptest database to test the ctfchallenge dataset 1. Amber ran ace, ace2,ctffind, phasor, and ctfRefine on them last week.
I looked at them today. Apart from ctfRefine, all the other seems to give the expected sign in the result plot, myamiweb, and database. I fixed the bug regarding ctfRefine and re-ran ctfRefine6 which gave similar results as others. The best image to look at is
zz13jul08z-ctfchal_set1_set_001_challenge_mic_006.mrc
It has a visually detectable astigmatism at -58 degrees.
makestack2.py result was checked with --fliptype=ace2image and --keepall. Keepall flag leaves the ctf-corrected image undeleted when the run is finished. Therefore, I could calculate from it the power spectrum and examine the astg_angle applied in the correction. It looks good, too (-58)
The use of astig_angle in apFrealign.py was checked by preparing single model reconstruction for Frealign and examining its parameter files. It looks good, too (-58).
Also checked parameter output for stack in myamiweb such as
/myamiweb/processing/listCTFvalues.php?expId=11956&sId=5
Remaining tasks:
1. Is there anything need changing for ctftilt? Neil?
2. Selection of bestCtfValue. Is it consistent with the display in image viewer? Amber?
3. Write dbschema update script. Anchi
4. Do we have xmipp result? Amber?
Updated by Anchi Cheng about 11 years ago
More question:
Amber, is there changes I might need to do for relion preparation?
Updated by Anchi Cheng about 11 years ago
One more for Neil,
Is inteactiveCTF ready for testing? We did not include it in the testing this.
Updated by Anchi Cheng about 11 years ago
- Status changed from New to In Code Review
- Assignee changed from Anchi Cheng to Neil Voss
r17982 changes all the signes
r17983 is the schema update related files for this change.
scan-ctfdata.py can be used to find if a user is still adding ctf values to the database or using the table within 1 day, so that its update need to be delayed.
Database must be updated if appion python/php codes are updated.
Updated by Anchi Cheng almost 11 years ago
The schema change was applied to NRAMM database 1 month ago and there is no report on problem resulting from it. Looks like a success.