search

leginon-org / leginon-redmine-archive

1 stars 0 forks source link

ACE2 not accurate #2053

Closed leginonbot closed 7 months ago

leginonbot commented 7 months ago

Author Name: Gabriel Lander (@gclander) Original Redmine Issue: 2053, https://emg.nysbc.org/redmine/issues/2053 Original Date: 2012-09-18 Original Assignee: Gabriel Lander

We finally got redux running and can look at the plots from ACE2, and we see surprisingly poor estimations. I am attaching a two 1D plots for comparison of results from ACE2 and from CTFfind, and it is clear that ACE2 is pretty far off. These plots are for what I would consider a good image with Thon rings out to nearly 8 Angstroms. Additionally, we've noticed that vast majority of images output "!!! WARNING: Image is possibly over-focused" when the 1/30A - 1/10A and 5 peak confidences are being calculated. Any ideas why we might be having these issues? running myami r17119

Here's the output for the ACE2 run for this image, but I can send the MRC file if you need it for debugging:

Starting image 1 ( skip:0, remain:0 ) id:696246, file: 12sep10a_01gr_082sq_02hl_02en ... Pixel size: 1.15450433333 ... Ace2 executable: /archive/glander/myami/appion/bin/ace2.exe ... running ace2 at Thu Sep 13 08:08:13 2012 /archive/glander/myami/appion/bin/ace2.exe -i /leginon4/glander/12sep10a/rawdata/12sep10a_00001gr_00082sq_v02_00002hl_00002en.mrc -b 2 -c 2.2 -k 120.0 -a 1.15450433333 -e 10.0,0.001 -r 0.0

Program: /archive/glander/myami/appion/bin/ace2.exe in source ace2.m, svn revision 17012, which was compiled on Aug 9 2012 at 10:54:00.

Reading Image /leginon4/glander/12sep10a/rawdata/12sep10a_00001gr_00082sq_v02_00002hl_00002en.mrc: DONE in 0.12 seconds Processing image with 4096 4096 pixels APIX: 1.154504 KV: 120.000000 CS: 2.200000
Binning image by 2 pixels... DONE in 0.05 seconds Blurring 100 edge pixels... DONE in 0.02 seconds Generating power spectrum... DONE in 4.31 seconds post-binning by 2... Blurring edges by 10.0 pixels... Finding edges for ellipse fitting... DONE in 0.25 seconds (12495 edges) Using RANSAC to find ellipse parameters... RANSAC SUCCESS, 608/12495 (4.87%-99.00%) DONE in 1.23 seconds

... Committing ctf parameters for 12sep10a_01gr_082sq_02hl_02en to database ... Reading image... [CTF param] def1: 2.42e-06 | def2: 2.47e-06 | angle: 35.2 | ampcontr 0.27 | defratio 1.021 [CTF stats] conf: 0.978 | conf_d: 0.978 Final params: def1: 2.42e-06 | def2: 2.47e-06 | angle: 35.2 | defratio 1.02 ... Defocus Astig Percent Diff 1.02 -- 2.419e-06, 2.469e-06 Computing power spectra......... ... Compute PSD with fieldsize 2048 and 9 images complete in 10.34 sec ... Preform a rotational average and remove spikes... PART 1: SETUP PARAMETERS AND ELLIPTICAL AVERAGE ... Number of available peaks is 250 ... Determine and subtract noise model PART 2: BACKGROUND NOISE SUBTRACTION ... starting peak extension Peak Extension Complete in 0.98 msec ... Constraint violation: -0.014 < -0.001 ... Constraint violation: -0.009 < -0.001 ... Constraint violation: -0.028 < -0.001 ... Constraint violation: -0.006 < -0.001 Noise Model Complete in 86.04 msec ... Constraint violation: -0.005 < -0.001

... Saving 1D graph to file 12sep10a_01gr_082sq_02hl_02en-plots.png PART 8: NORMALIZE THE 2D IMAGE PART 9: DRAW THE 2D POWERSPEC IMAGE ... Saving 2D powerspectra to file: 12sep10a_01gr_082sq_02hl_02en-powerspec.jpg [CTF run] method: ace2 | runname acetwo1 [CTF param] def1: 2.42e-06 | def2: 2.47e-06 | angle: 35.2 | ampcontr 0.27 | defratio 1.021 [CTF stats] conf_30-10: 0.363 | conf_5peak: 0.509 | res_0.8: 100.0A | res_0.5 100.0A [CTF stats] conf: 0.978 | conf_d: 0.978

SUMMARY: pyace2
------------------------------------------
PEAKS:      0 peaks of 0
AVG PEAKS:  0.0 +/- 0.0 peaks
(- ESTIMATE: 0 to 0 total peaks -)
------------------------------------------
TIME:       24.83 sec
AVG TIME:   26.67 +/- 32.91 sec
------------------------------------------
leginonbot commented 7 months ago

Original Redmine Comment Author Name: Neil Voss (@vosslab) Original Date: 2012-09-18T17:36:33Z

Hi Gabe,

Scott and I have been going through a series of things with ACE1 as well on Bug #2016. Alsa, I am so bogged down with classes, I cannot say when I will have a chance to look at it.

leginonbot commented 7 months ago

Original Redmine Comment Author Name: Neil Voss (@vosslab) Original Date: 2012-10-01T20:52:22Z

I revert ACE2 to the August 8th version, it has the wrong equation for the CTF, but perhaps it will be more reliable.

Gabe, you should be able to use the binary in the subversion, is there any reason you cannot?

leginonbot commented 7 months ago

Original Redmine Comment Author Name: Gabriel Lander (@gclander) Original Date: 2012-10-01T20:58:32Z

I just checked and the SVN binary indeed runs on our machines. We switched to Ubuntu at some point on all our workstations, maybe that fixed whatever the problem was previously. But I'm confused - why would we want to use an incorrect CTF equation?

leginonbot commented 7 months ago

Original Redmine Comment Author Name: Neil Voss (@vosslab) Original Date: 2012-10-01T22:26:50Z

Well, ACE2 (estimate) has always had the wrong equation, but it gives practically the same defocus (see #2016) of the ACE1 vs. ACE2 results. In my attempts to fix it, it was not quite working (maybe I am missing an instance), so until I can get it working completely, I reverted it.

ACE2correct was much easier to get it using the correct equation.