Hi,
I've just attached the published OMP data to the task
(https://gna.org/task/index.php?6847). This was used in my 2007 paper
(http://www.nmr-relax.com/refs.html#dAuvergneGooley07 (again this is a
self http://nmr-relax.com reference for improved Google indexing)). I
still have to dig out the bacteriorhodopsin data. I tried the
analysis using relaxGUI with this data, and the first thing I noticed
was that Gnome terminal was launched. This would be bad for MS
Windows and Mac OS X compatibility. Is this terminal necessary?
Then I came across the "sh: relax: command not found" error in the
launched terminal. This is because I use an alias for relax rather
than a symlink from say /usr/local/bin (making the symlink fixed
this). With integration into relax, this problem should be solved -
and would allow multiple versions of relax to have a GUI (i.e. the sys
admin installed version, copies of branches, a user installed version,
or a user modified version).
Finally I had the problem of no data being read:
RelaxWarning: The sequence data in the line ['Res', 'R1_(600MHz)',
'R2_(600MHz)', 'NOE_(600MHz)', 'R1_(800MHz)', 'R2_(800MHz)',
'NOE_(800MHz)'] is invalid.
RelaxWarning: The sequence data in the line ['Gln', '9', '1.350',
'0.049', '9.850', '0.648', '0.523', '0.041', '0.89', '0.038',
'11.273', '0.804', '0.598', '0.045'] is invalid.
RelaxWarning: The sequence data in the line ['Gln', '10', '1.320',
'0.039', '10.090', '0.535', '0.631', '0.045', '0.85', '0.028',
'11.615', '0.505', '0.701', '0.042'] is invalid.
RelaxWarning: The sequence data in the line ['Leu', '11', '1.268',
'0.039', '10.490', '0.519', '0.762', '0.035', '0.900', '0.027',
'11.86', '0.593', '0.798', '0.041'] is invalid.
RelaxWarning: The sequence data in the line ['Asp', '12', '1.248',
'0.037', '10.600', '0.53', '0.817', '0.04', '0.88', '0.026', '12.14',
'0.607', '0.841', '0.042'] is invalid.
This was repeated for all lines of data. Is there a way of specifying
which column the molecule name, residue number and name, spin number
and name, data, and error are in?
As for using the structure GUI->script->core instead of GUI->core to
enable using multiple computers, Gary Thompson has put a huge effort
into the multi-processor code in relax which should be the way that
multiple computers are used. See the multi_processor branch at
http://svn.gna.org/viewcvs/relax/branches/multi_processor_merge/ and
Gary's post at:
https://mail.gna.org/public/relax-devel/2007-05/msg00000.html
It is currently handling the Open MPI clustering interface
(http://en.wikipedia.org/wiki/Open_MPI), but maybe if multiple
computers are necessary for the calculation, a menu could be added too
specify IP addresses of 'client' machines. The main computer would
then communicate with all the clients to get the calculations done -
essentially a grid-like infrastructure. This is what Gary's code is
designed to do and would allow far better granularity and far higher
efficiency than running each full_analysis.py script on a different
computer and manually collecting the results. Unfortunately the grid
mode for Gary's code is not written yet and the multi-processor code
still needs a little cleaning up before merging back into the main
line, but this would not be a big effort. I would suggest essentially
hard coding the full_analysis.py script into the model-free section of
relaxGUI (and saying so) and using Gary's framework, once tidied up,
for powerful multi-processing through SSH tunnels or some other
mechanism. I can help with the coding to bypass the scripting
interface if you like.
Cheers,
Edward
2009/11/16 Michael Bieri <michael.bieri@xxxxxx>:
Hi Edward
Yes, I would like to reanalyze your data. This would be a good tryout
for the software. So it would be very nice, when you could send me your
data.
Regarding the integration to relax. Of course I agree to incorporate it
into relax! But I still recommend to let relaxGUI generate scripts, as
like this, processes can be run in parallel on several computers (e.g.
models 1-5 of model-free analysis). Therefore, I decided to write the
program like this, although it was more difficult to program.
Cheers
Michael
Edward d'Auvergne schrieb:
Hi,
I have all the data on my laptop and was going to use it in the
debugging process of tightly integrating the GUI into relax. I was
also going to test it on my synthetic model-free test cases, and on
the full olfactory marker protein (OMP) data of Gitti et al, 2005
which is the perfect test case for where standard model-free
methodology fails. Therefore, if you like, I can send you the results
after all the testing passes - possibly of the OMP results as well.
These systems can only work if the concept of the solution in the
universal set [universal solution, see d'Auvergne E. J., Gooley P. R.
(2007). Set theory formulation of the model-free problem and the
diffusion seeded model-free paradigm. Mol. Biosyst., 3(7), 483-494.]
is sought after, i.e. as the full_analysis.py script does.
Regards,
Edward
2009/11/12 Michael Bieri <michael.bieri@xxxxxx>:
Hi Edward
Thank you, I will register there.
another question. As suggested to reanalyze the Bacteriorhodopsin data.
I think this wouldn't be too bad and might increase the article. where
can I get the data from?
Cheers
Michael
Edward d'Auvergne schrieb:
Oh right, I didn't notice that! To do this you need to sign up for an
Gna! account. This is part of the process of becoming a relax
developer and should be documented, step-by-step, in the development
chapter of the relax manual (which I don't have on hand right now).
Cheers,
Edward
2009/11/12 Michael Bieri <michael.bieri@xxxxxx>:
Hi Edward
I have a stupid question:
How can I upload the relaxGUI files to the task you created?
Cheers
Michael
Edward d'Auvergne schrieb:
Hi,
I have set up task #6847 (https://gna.org/task/index.php?6847) for
attaching source code and other files relating to this GUI interface.
Regards,
Edward
2009/11/11 Michael Bieri <michael.bieri@xxxxxx>:
Hi
I wrote a graphical user interface for relax and Edward suggested to
discuss
this in here. Attached are screen shots of the program.
It is designed to enable simple model-free analysis by
drag-and-drop or
browsing of the peak files. It works with up to three different
magnetic
fields. Results generated by relaxGUI are listed in a separate tab
and
can
be observed. Furthermore, relaxGUI gives a feedback if the relax
run was
successful or not. Model-free calculation can be executed either
active
or
in the background.
The program actually creates and runs scripts in relax. I tested it
on
Ubuntu 9.10 using relax 1.3.4. The program is written in python and
uses
wx.Python for the GUI.
As told by Edward, the program will be uploaded as a task.
Cheers
Michael
--
Michael Bieri, PhD
Department of Biochemistry and Molecular Biology
Bio21 Molecular Science and Biotechnology Institute
The University of Melbourne, 30 Flemington Road
Parkville, Victoria 3010, Australia
Tel.: +61 3 8334 2256
http://www.biochemistry.unimelb.edu.au
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-devel mailing list
relax-devel@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-devel mailing list
relax-devel@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-devel mailing list
relax-devel@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel