user queries for the Xplor-NIH biomolecular structure determination program

headers Andrew Severin | 6 Sep 2007 22:25

_______________________________________________
Xplor-nih mailing list
Xplor-nih <at> nmr.cit.nih.gov
http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih

Hello Andrew--

> I have noticed when I use this script with my data that I get this
> 
> WARNING: large difference in mode (maximum of distribution)
> calcDaRh.py(385):     print "  observed: %7.3f  calculated: %7.3f" %
> (modeObs,modePowder)
>   observed:   5.920  calculated: -19.595
> 
> 
> I realize that this means the calculated result isn't representative of the
> data.  Does this mean I can not use this script with my data?  Is there a
> common error I may have made with the data that causes this error?

I think there's a bug in the released version of calcDaRh.py. Please
grab an updated version from 
http://nmr.cit.nih.gov/xplor-nih/updates/current/

The warning may go away with the new version.

thanks--
Charles

Hello!

I have been trying to refine the structure of my protein using dipolar 
coupling constraints using the latest version (2.18) of Xplor-NIH. I 
downloaded and installed the package for both Linux (CentOS) and Mac (iMac).

When I ran my data files, I found that the output files showed different 
violations for the same input files in the two OSs. I used the same 
seed, the same set of dipolar couplings and the same input structure, 
for refinement. The iMac version shows more violations in the output 
file vs the one installed in the CentOS system. Has this been seen 
before? Which output file do I work on?
I can mail the input and output files if you want to have a look at it.

Thank you so much!

With best regards,
Maha
---
R. Mahalakshmi
Postdoctoral Associate
Marassi Lab
Burnham Institute for Medical Research
La Jolla, CA 92037
Ph: 858-646-3100x3724

Hello Maha--

> When I ran my data files, I found that the output files showed different 
> violations for the same input files in the two OSs. I used the same 
> seed, the same set of dipolar couplings and the same input structure, 
> for refinement. The iMac version shows more violations in the output 
> file vs the one installed in the CentOS system. Has this been seen 
> before? Which output file do I work on?

The two binaries were produced with different compilers and utilize
different libraries on the two platforms- therefore, different results
are expected. However, when you calculate enough structures, these
differences should wash out. 

If you are convinced there is a problem, please calculate 20 structures
on each platform. If the results in the resulting .stats files do not
overlap, then there probably really is a problem, and I should take a
look.

best regards--
Charles

Hi
I am using the Fit and RMSD objects to fit structures to a starting 
structure as shown in the example python/tests/fitTest.py . However, it 
seems to be incredibly slow (it takes minutes to fit 50 structures with 
CA overlay and 150 residues...) is this expected

regards
gary

--

-- 
-------------------------------------------------------------------
Dr Gary Thompson
Astbury Centre for Structural Molecular Biology,
University of Leeds, Astbury Building,
Leeds, LS2 9JT, West-Yorkshire, UK             Tel. +44-113-3433024
email: garyt <at> bmb.leeds.ac.uk                   Fax  +44-113-2331407
-------------------------------------------------------------------

Hi Gary,

Sounds painful.  The TCL version is very fast.  Take a look at
xplor/tcl/noe_reporting.tcl  There's a proc called  
summarizeMarvinResults.
Inside it are calls to TCL procs called meanStruct,  
compareToReference, and structPrecision.

You'd use it something like this:

set allFiles [grabPDBFiles -fileNames [glob *.pdb]]

set backboneSel [AtomSel -args "(name ca or name c or name n)"]
set heavyatomSel [AtomSel -args "(not name h*)"]

meanStruct -files $allFiles -selection $backboneSel

set bbnRMSD [compareToReference -referenceFileName "answer.pdb" - 
selection $backboneSel]
set heavyRMSD [compareToReference -referenceFileName "answer.pdb" - 
selection $heavyatomSel]

set precisionReport [structPrecision -files $allFiles \
					-backboneSelection $backboneSel \
					-heavyatomSelection $heavyatomSel \
					-verbose]

puts [format "Accuracy of the mean structure is %f A backbone, %f A  
heavyatoms" $bbnRMSD $heavyRMSD]
puts [format "Precision of the individual ensemble members to the  
mean is %s" $precisionReport]

Hope this helps.

--JK

On Sep 27, 2007, at 9:27 AM, Gary Thompson wrote:

> Hi
> I am using the Fit and RMSD objects to fit structures to a starting
> structure as shown in the example python/tests/fitTest.py .  
> However, it
> seems to be incredibly slow (it takes minutes to fit 50 structures  
> with
> CA overlay and 150 residues...) is this expected
>
> regards
> gary
>
> -- 
> -------------------------------------------------------------------
> Dr Gary Thompson
> Astbury Centre for Structural Molecular Biology,
> University of Leeds, Astbury Building,
> Leeds, LS2 9JT, West-Yorkshire, UK             Tel. +44-113-3433024
> email: garyt <at> bmb.leeds.ac.uk                   Fax  +44-113-2331407
> -------------------------------------------------------------------
>
>
> _______________________________________________
> Xplor-nih mailing list
> Xplor-nih <at> nmr.cit.nih.gov
> http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih

Hello Gary--

> I am using the Fit and RMSD objects to fit structures to a starting 
> structure as shown in the example python/tests/fitTest.py . However, it 
> seems to be incredibly slow (it takes minutes to fit 50 structures with 
> CA overlay and 150 residues...) is this expected

That is not expected. Something's wrong- I suspect an I/O issue. Which
platform are you running on? Is you send your script I should be able to
help.

regards--
Charles

> Hello Gary--
>
>> > I am using the Fit and RMSD objects to fit structures to a starting 
>> > structure as shown in the example python/tests/fitTest.py . However, it 
>> > seems to be incredibly slow (it takes minutes to fit 50 structures with 
>> > CA overlay and 150 residues...) is this expected
>
> That is not expected. Something's wrong- I suspect an I/O issue. Which
> platform are you running on? Is you send your script I should be able to
> help.
>
> regards--
> Charles

Hi Charles

IO speed could be the problem here as we do have known IO speed problems its just that i wasn't aware that they
were that bad, I will have to investigate further

regards
gary
--

-- 
-------------------------------------------------------------------
Dr Gary Thompson
Astbury Centre for Structural Molecular Biology,
University of Leeds, Astbury Building,
Leeds, LS2 9JT, West-Yorkshire, UK             Tel. +44-113-3433024
email: garyt <at> bmb.leeds.ac.uk                   Fax  +44-113-2331407
-------------------------------------------------------------------

_______________________________________________
Xplor-nih mailing list
Xplor-nih <at> nmr.cit.nih.gov
http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih

Hello Gary--

> 
> IO speed could be the problem here as we do have known IO speed
> problems its just that i wasn't aware that they were that bad, I will
> have to investigate further 
> 

For reading the files in the loop: use pdbTool, not the helpers in
protocol, as the former has less overhead. You also might try
eginput/gb1_rdc/aveStruct.py, as that fits structures as a side-effect.

beest regards--
Charles

11	May 2013
21	April 2013
7	March 2013
4	February 2013
8	January 2013
1	November 2012
4	October 2012
2	September 2012
2	August 2012
5	July 2012
11	June 2012
4	April 2012
25	March 2012
6	February 2012
3	January 2012
3	December 2011
14	November 2011
7	October 2011
9	September 2011
4	August 2011
4	July 2011
7	June 2011
6	May 2011
5	April 2011
5	March 2011
6	February 2011
21	January 2011
16	December 2010
17	November 2010
27	October 2010
15	September 2010
5	August 2010
13	July 2010
36	June 2010
9	May 2010
9	April 2010
14	March 2010
27	February 2010
12	January 2010
17	November 2009
8	October 2009
17	September 2009
16	August 2009
33	July 2009
27	June 2009
17	May 2009
29	April 2009
11	March 2009
23	February 2009
17	January 2009
2	December 2008
25	November 2008
15	October 2008
26	September 2008
17	August 2008
21	July 2008
15	June 2008
26	May 2008
61	April 2008
19	March 2008
23	February 2008
16	January 2008
45	December 2007
20	November 2007
20	October 2007
11	September 2007
46	August 2007
29	July 2007
38	June 2007
29	May 2007
34	April 2007
33	March 2007
21	February 2007
7	January 2007
16	December 2006
22	November 2006
27	October 2006
10	September 2006
29	August 2006
16	July 2006
15	June 2006
20	May 2006
20	April 2006
13	March 2006
18	February 2006
7	January 2006
25	December 2005
11	November 2005
6	October 2005
2	September 2005
4	August 2005
10	July 2005
5	June 2005
6	May 2005
3	April 2005
4	March 2005
1	January 2005
4	November 2004
7	October 2004
8	September 2004
10	August 2004
6	July 2004
13	June 2004
5	May 2004
9	April 2004
10	March 2004
14	February 2004
10	January 2004
7	December 2003
15	November 2003
10	October 2003
6	September 2003
2	August 2003
12	July 2003
4	June 2003
8	May 2003
2	April 2003
6	March 2003
6	February 2003
8	January 2003
4	December 2002
5	October 2002
1	September 2002
7	August 2002
8	July 2002
6	June 2002

calcDaRh.py

Re: calcDaRh.py

Different outputs for different OSs

Re: Different outputs for different OSs

fitting speed

Re: fitting speed

Re: fitting speed

Re: fitting speed

planeDistPot

Re: fitting speed

Mon	Tue	Wed	Thu	Fri	Sat	Sun

					1	2
3	4	5	6	7	8	9
10	11	12	13	14	15	16
17	18	19	20	21	22	23
24	25	26	27	28	29	30