search

VuisterLab / cing

Automated Validation of NMR Structures
http://nmr.le.ac.uk
2 stars 4 forks source link

Chemshifts of methyl protons missing from CCPN project when imported. #192

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1.From Vitaliy's mail:

After the transfer of my CCPN project I have found that chemical shifts of all
methyl protons are absent in CING results.

He send me the project and the CING results which I can't post because they're 
unpublished.
Most likely, the problem has to do with the warning:

WARNING: This routine wasn't meant to be used for atoms that are part of a 
group of more than 2; please improve code

Original issue reported on code.google.com by jurge...@gmail.com on 9 Oct 2009 at 9:03

GoogleCodeExporter commented 9 years ago 
I tried using CCPN Analysis for reading Vitaliy's project but got the attached 
message:

No module named downgrade.v_2_0_b3.

I tried updating CCPN Analysis but there were no updates.

CING reads the project fine (other than this issue) so ...

Tim, is there another way to update Analysis to use the latest CCPN API? You 
might have told me before...

I am also checking another CCPN project with methyl CS so I'm not stuck on this 
issue.
http://code.google.com/p/cing/source/browse/trunk/cing/Tests/data/ccpn/1brv_cs_p
k_2mdl.tgz

It looks like there is a problem with that one too:
https://nmr.cmbi.ru.nl/tmp/cing/ano/S0ONWy/1brv_cs_pk_2mdl.cing

Original comment by jurge...@gmail.com on 12 Oct 2009 at 10:11

Attachments:

GoogleCodeExporter commented 9 years ago 
Tim, Analysis also can't open the uncompressed:
http://code.google.com/p/cing/source/browse/trunk/cing/Tests/data/ccpn/1brv_cs_p
k_2mdl.tgz

Same bug.
jd:Stella/cingTmp/ analysis 1brv_cs_pk_2mdl
Model read finished. Duration 2.05983805656 
Model validity check skipped
CcpNmr Analysis Version 2.0. Release 7 (Copyright 2003-2009 CCPN)
Distribution created Mon Apr  6 16:53:39 2009

If you use this software, please quote the following reference:

Wim F. Vranken, Wayne Boucher, Tim J. Stevens, Rasmus H. Fogh, Anne 
Pajon, Miguel Llinas, Eldon L. Ulrich, John L. Markley, John Ionides
and Ernest D. Laue (2005). The CCPN Data Model for NMR Spectroscopy: 
Development of a Software Pipeline. Proteins 59, 687 - 696.

The FAQ is at: http://www.ccpn.ac.uk/ccpn/software/ccpnmr-analysis/faq
For further help see the mailing list: http://www.jiscmail.ac.uk/ccpnmr

No compatibility module found for version 2.0.b3
Error loading file for:  None
Reading:  <open file 
'/Users/jd/tmp/cingTmp/1brv_cs_pk_2mdl/memops/Implementation/1brv.xml', mode 
'r' 
at 0x4a689b0>
Last xml tag read:  _StorageUnit
Parser state was:  starting
Object stack was empty
No compatibility module found for version 2.0.b3
Error loading file for:  <memops.Implementation.MemopsRoot id:100062288>
Reading:  <open file 
'/Users/jd/tmp/cingTmp/1brv_cs_pk_2mdl/memops/Implementation/1brv.xml', mode 
'r' 
at 0x5f518d8>
Last xml tag read:  IMPL.MemopsRoot.packageLocators
Parser state was:  handling version compatibility
Current object was:  <memops.Implementation.MemopsRoot id:100062288>
Traceback (most recent call last):
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/ccpnmr/analysis/AnalysisGui.py", line 224, in 
<module>
    main(projectDir, max_size, glDirect)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/ccpnmr/analysis/AnalysisGui.py", line 102, in main
    project = loadProject(top, path=projectDir)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/ccp/gui/Io.py", line 81, in loadProject
    askDir=askdir, askFile=askfile)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/general/Io.py", line 531, in loadProject
    project = XmlIO.loadProject(path, projectName, partialLoad=True)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/format/xml/XmlIO.py", line 205, in 
loadProject
    return loadProjectFile(filePath, partialLoad=partialLoad)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/format/xml/XmlIO.py", line 217, in 
loadProjectFile
    partialLoad=partialLoad)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/format/xml/XmlIO.py", line 267, in 
loadFromStream
    partialLoad=partialLoad)
  File "/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/xml/Implementation.py", line 4778, in 
loadFromStream
    Converters1.minorPostProcess(fileVersion, result, delayDataDict, objectDict, mapping)
  File 
"/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/format/compatibility/part1
/Converters1.py", 
line 894, in minorPostProcess
    mod = getCompatibilityModule(fromVersionStr)
  File 
"/Users/jd/workspace34/ccpnmr/ccpnmr2.0/python/memops/format/compatibility/part1
/Converters1.py", 
line 869, in getCompatibilityModule
    result = __import__(moduleDir, {}, {}, compModules)
ImportError: No module named downgrade.v_2_0_b3

Now, I'm stuck because I can't 'see' if the CS are actually in the CCPN project.

Original comment by jurge...@gmail.com on 12 Oct 2009 at 10:16

GoogleCodeExporter commented 9 years ago 
From Wayne:

I think the problem is almost certainly with your Analysis profile (stored in 
~/.ccpn/data/ccpnmr/AnalysisProfile/SOMETHING_HORRID.xml).  So it is a file 
that is shared across multiple Analysis 
versions.  And if you ever save a project in Analysis 2.1 then the release in 
that file will be set to 2.0.b3.  So from then on you cannot use Analysis 2.0 
because it has to be able to read that Analysis 
profile but cannot because it is a newer version.  That is what that opaque 
message is saying.  So as a quick work-around (which we have to do in 
Cambridge, so we know how irritating it is) edit 
that file and change the release to 2.0.b2 (in the second line).  But you have 
to do that every time you save something in 2.1 and then go back to using 2.0.  
Rasmus has added some ability to do 
forwards compatibility which should get around this problem, but it hasn't been 
distributed to the world yet.  (I guess he could probably ship it to you as a 
guinea pig.)

Removing the entire .ccpn directory helped as a one-time work-around.

Thanks Wayne

Now I can continue with the issue.

Original comment by jurge...@gmail.com on 14 Oct 2009 at 9:35

GoogleCodeExporter commented 9 years ago 
Confirming this problem as observed for test data:
http://code.google.com/p/cing/source/browse/trunk/cing/Tests/data/ccpn/1brv_cs_p
k_2mdl.tgz

I had to update it to actually add the cs in there. The Leu 185 methyl CS are 
present in the CCPN project but not 
in the CING report. Other CS do carry over.

Original comment by jurge...@gmail.com on 14 Oct 2009 at 10:16

GoogleCodeExporter commented 9 years ago 
# Since we don't show methyls in the assignment list any more; they appear 
lost; this fixes issue 192.
Committed in r645. Updated CING server to r645 so Vitaliy can check for his 
large project.

I checked for 1brv example to:
https://nmr.cmbi.ru.nl/tmp/cing/ano/BdCTFy/1brv_cs_pk_2mdl.cing/1brv/HTML/Molecu
le/atoms.html#_top
leu 185 QD1 is there.

Original comment by jurge...@gmail.com on 14 Oct 2009 at 12:09

GoogleCodeExporter commented 9 years ago 
From V.:

The issue with methyls is not completely solved, unfortunately. The two groups 
of Val and Leu methyl carbons 
and protons have the same chemical shifts in CING results while in the CCPNMR 
project they have got their 
chemshifts different, e.i. their NMR signals do not overlap.

Original comment by jurge...@gmail.com on 19 Oct 2009 at 8:45

GoogleCodeExporter commented 9 years ago 
I'm stuck again because the trick with removing the ~/.ccpn directory fails for 
V.'s anonymous project file with 
my Analysis.
Can you provide me with an older version of a CCPN project or any other project 
where this happens. As shown 
in the attached gif it works fine for 1brv A    183 LEU QD1 and QD2
at:
https://nmr.cmbi.ru.nl/tmp/cing/ano/BdCTFy/1brv_cs_pk_2mdl.cing/1brv/HTML/Molecu
le/atoms.html#_top

Original comment by jurge...@gmail.com on 20 Oct 2009 at 9:40

Attachments:

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
That one is from oct 15. It needed to rerun it with a CING version r645 or 
later as mentioned in comment 5.
Could you do that?

So you don't mind the data being public here? Otherwise remove your previous 
comment.

Original comment by jurge...@gmail.com on 20 Oct 2009 at 11:46

GoogleCodeExporter commented 9 years ago 
How can I remove my project from the server?
How can I delete the comment?
Thank you.

Original comment by vya...@gmail.com on 20 Oct 2009 at 3:29

GoogleCodeExporter commented 9 years ago 
Jurgen,
The header of cingRun.log of the above CING run for my project says r645. So I 
used
the newest revision on oct 15 and it gave the identical chemshifts for methyl 
groups
of Val and Lue.
Also, I have started the new CING run to double check that the issue remains.

In CING window of CCPN package I am not able to send my project using https. It 
gives
error about unknown url. But it works with http. Is https not implemented yet?
V

Original comment by vya...@gmail.com on 20 Oct 2009 at 4:49

GoogleCodeExporter commented 9 years ago 
The new CING run on my project started from CCPN confirms the above mentioned
problems with CH3 assignments for Val and Leu

I have noticed another issue. I selected in CING window of the CCPN project to 
use in
CING validation a set of structures and one chemical shift list. The CCPN 
project has
also peaks and restraints which I did not add to be used in validation. However 
they
all have been picked up and validated as it is seen in CING results.

Original comment by vya...@gmail.com on 20 Oct 2009 at 8:10

GoogleCodeExporter commented 9 years ago 
-1- Tim, can you confirm the trouble vyagor is seeing with the https and 
selection not working? From my end icing is working fine with https.

-2- V. can you upload here a different small public project with the same CH3 
CS problem? I'm stuck on this issue as I can't load your project into my 
Analysis code to check and see what the input to CING is.

-3- V., to delete the data on our server you can use the Analysis or iCing 
interfaces or you can programmatically do it using the code in:
http://code.google.com/p/cing/source/browse/trunk/cing/python/cing/Scripts/iCing
Robot.py

-4- V. to delete a comment here, look at the right hand side all the way over 
and there may be the text "Delete comment". If there isn't then you might not 
have permissions to do so. Anyway, I already removed 
your comment.

Original comment by jurge...@gmail.com on 21 Oct 2009 at 10:28

GoogleCodeExporter commented 9 years ago 
Unfortunately I do not have another project that I could make public...
But 1brv example above also has this problem. In Analysis I can see that L168 
Hda*
has 0.900ppm and Hdb* has 0.930ppm. In the Cing results they have 0.90ppm. 
Similar
happens for L180 Hda* and Hdb* (carbons have the same chemshift in the 
Analysis).
Similar happens for L183 Cda, Cdb but NOT for their protons. Methyl protons of 
L183
in Analysis are assigned as Hd1* (0.940ppm) and Hd2* (0.760ppm). They got 
transferred
to CING results correctly. The two methyls of V171 are assigned as Hg1* and 
Hg2* and
have the same chem shift in the project. I have intentionally changed the chem 
shift
for V171 Hg2* to some other value and it has appeared in the CING results. So I 
guess
it is the reason: a* and b* can make it while for 1* and 2* it works.

How do I use iCingRobot.py?

Original comment by vya...@gmail.com on 21 Oct 2009 at 7:31

GoogleCodeExporter commented 9 years ago 
-1- I just checked the https submit in Analysis and it works fine. Are you 
still having problems with it?

-2- In answer to your comment 14:
  >  L168 Hda* has 0.900ppm and Hdb* has 0.930ppm. In the Cing results they have 0.90ppm. 
  I can see that happening because we might not do the resonance concepts perfectly in CING yet.
I think you mean to conclude instead of:
>So I guess it is the reason: a* and b* can make it while for 1* and 2* it 
works.
that:
>So the difference between a* and b* is not maintained in the importer from 
CCPN to CING while it is for 1* and 2*.
I'm leaving this issue open for us to address it later.

-3- There's a unit check for it that you can study and execute to learn the 
robot. We do it every time before committing to svn.
python -u $CINGROOT/python/cing/Scripts/test/testiCingRobot.py
You would overide the access_key variable with the one you would want to 
remove. It should also demonstrate that https is working. For that you need the 
r649 in which I've set 
the default to be:
DEFAULT_URL = 'https://nmr.cmbi.ru.nl' # production with https security turned 
on

Thanks a lot vyagor, it's great to have outside input like yours!

Original comment by jurge...@gmail.com on 22 Oct 2009 at 8:29

GoogleCodeExporter commented 9 years ago 
1. I have just checked again the submission from Analysis with https and have 
got:
"uknown url type: https"

2. Sorry for my unclear sentence. I meant "a* and b* can make it (the problem)"

3.I guess if I want to use iCingRobot.py script I need to install Cing on my
computer. I'll do it later. But now could you please delete my project.

Thank you,
Vitaliy

Original comment by vya...@gmail.com on 23 Oct 2009 at 1:59

GoogleCodeExporter commented 9 years ago 
The https problem is made into issue 202.
I have removed your project manually. Did you find the functionality in 
Analysis for the future use?

Cheers

Original comment by jurge...@gmail.com on 23 Oct 2009 at 8:03

GoogleCodeExporter commented 9 years ago 
Thank you!
Yes, Analysis has the button "Purge Server Result". I did not apply it for my 
project
results since I wanted you to have a look at my CING results regarding this 
methyl
chemshifts issue...

Cheers

Original comment by vya...@gmail.com on 23 Oct 2009 at 12:07

GoogleCodeExporter commented 9 years ago

Original comment by jurge...@gmail.com on 5 Nov 2009 at 1:08

GoogleCodeExporter commented 9 years ago 
So, in conclusion CING still has a problem with maintaining the difference 
between eg. 
> L168 Hda* has 0.900ppm and Hdb* has 0.930ppm. In the Cing results they have 
0.90ppm.

Ok, both have 0.93 now so this is still a valid issue.

Original comment by jurge...@gmail.com on 8 Feb 2010 at 1:53

GoogleCodeExporter commented 9 years ago 
Fixed recently.

Original comment by jurge...@gmail.com on 3 Jan 2011 at 9:46