Dear Christina Möller.
I am sure Edward can give you an answer to this.
I know that he is on holiday until the 7th, so he will probably first
respond after this.
Until then, I can try to give you some temporary idea's you can work with.
Beforehand, I would like to say, that I no experience with the whole setup
of modelfree, but
have only worked with the dispersion module of relax.
What you want to play with, is the function "value.set" and "minimise" in
relax.
Try this.
# Start relax in terminal mode
relax
# Then see help functions
help(minimise)
# Function to look into
help(minimise.grid_search)
help(minimise.execute)
help(minimise.calculate)
"minimise.grid_search" is the function to find initial values, before
minimise.execute
is performed.
Grid searching is "expensive", and particular if clustering of residues is
part of this.
I have tried to redo some published results. Often these results for
dispersion only gives the global fitted paramter
of exchange rate or population, but neglect to report the individual
chemical shift change (dw).
See http://wiki.nmr-relax.com/CR72.
What I can do is then:
value.set(val=2000.0, param='kex', error=False, force=True)
value.set(val=0.98, param='pA', error=False, force=True)
minimise.grid_search()
minimise.grid_search will detect that kex and pA has already been set, and
therefore then lock these, and
will then only search for the optimal "dw" and "r2" parameter.
After this, I will do:
minimise.execute()
But here, all parameters are "set free", when minimising, and I don't think
you can "lock" a parameter in this.
And the results are often "close" to the published values, and within
experimental error.
What you can also do, is just manually setting a parameter value and do:
minimise.calculate
minimise.calculate simply just takes the stored parameter values and
calculate the chi2 value.
So here, you could test a range of your own tm, and do minimise.calculate
to see the chi2 values.
This goes fine for maybe testing one or two ranges of values of a parameter.
You would probably like to make tm/chi2 graph, to map the change in chi2
value.
What is more powerful is to use the dx.map function.
http://wiki.nmr-relax.com/Install_dx
http://wiki.nmr-relax.com/Dx_map
This maps 3 parameters, to visualize the chi2 hypersurface.
See Edwards paper on this:
http://www.nmr-relax.com/refs.html
Model-free model elimination: A new step in the model-free dynamic analysis
of NMR relaxation data
http://link.springer.com/article/10.1007%2Fs10858-006-9007-z
This will give you a VERY clear idea, how the chi2 surface is mapped for
the tm values.
relax has a lot of user functions to help with analysis, and a lot more
"plumbing functions", to do exactly as you wish.
Here are some ideas for a start, but adjust to your example.
(This is for dispersion)
----------
state.load("final_state_from_your_analysis.bz2")
# A plumping function
from pipe_control.mol_res_spin import spin_loop
# Just test some print things
for spin, mol_name, resi, resn, spin_id in spin_loop(full_info=True,
return_id=True, skip_desel=True):
print dir(spin), spin.params
# Make an array
kex_arr = [1000., 2000., 3000.,]
for kex in kex_arr:
value.set(val=kex, param='kex', error=False, force=True)
minimise.calculate()
print kex, cdp.mol[0].res[0].spin[0].chi
-----
If you need specific help with a saved state for practicing, you can ask
for support.
http://gna.org/support/?group=relax
Here you can describe the problem, and add a saved state from your final
minimization.
Then I can write a little script to help you out.
Note, you should only add a state, where you have deleted all residues,
except 2-3.
This is make sure, not to compromise any confidentiality of your data.
Best
Troels
2015-06-03 14:32 GMT+02:00 Christina Möller <c.moeller@xxxxxxxxxxxxx>:
Dear Edward and relax users,
I successfully performed the dauvergne_protocol.py analysis scripts to
determine the ps - ns dynamics of a 14 kDa protein. The global
correlation time tm is 9.6 ns. Since other methods suggested a smaller
global correlation time tm, I would like to know whether it is possible
to fix the tm value in all rounds of individual model-free
optimisations? The corresponding chi2 values might then be useful to
evaluate the global correlation times that I obtained by different methods.
Best regards,
Christina
--
------------------------------------------------------------
Christina Möller
ICS-6 / Structural Biochemistry
Forschungszentrum Jülich
D-52425 Jülich
GERMANY
E-mail: c.moeller@xxxxxxxxxxxxx
Tel.: +49-2461-619387
Fax +49-2461-619387
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),
Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
_______________________________________________
relax (http://www.nmr-relax.com)
This is the relax-users mailing list
relax-users@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-users
Re: fixed tm values