Re: scripts -- January 03, 2007 - 11:02

Hi Hongyan,

relax is designed to be completely flexible in the way you perform your
analysis, allowing for the procedure to be tailored to the system at
hand, or for new proceedures to be developed. One procedure that I can
recomend which is somewhat similar to the one you outline is as follows:

1. estimate tm
2. fit each residue to dynamic models
3. select best model
4. fit tm and dynamic parameters simultaneously
5. repeat steps 2-4 starting from best-fit tm value. Continue until
results converge
6. repeat steps 2-5 for each diffusion model (isotropic, axially
symetric and anisotropic)
7. select best diffusion model
8. Monte Carlo simulations (error analysis)

As you note, Monte Carlo simulations over all parameters will be very
slow. This is why I recommend only performing the error analysis at the
end of the whole proceedure. I some cases it may be necessary to perform
the Monte Carlo simulations over only the dynamic parameters (ie. with
diffusion tensor fixed) in order to improve efficiency.

There has been some discussion of this and other analysis proceedures on
this list before. The thread that starts here:

https://mail.gna.org/public/relax-users/2006-10/msg00007.html

is worth a look.

Chris



On Wed, 2007-01-03 at 11:45 +0800, Hongyan Li wrote:
> Dear Chris,
>
> Thanks for the help! Can I do both i.e. Minimise all parameters and Monte 
> Carlo
> Simulations? I am still a bit confused the procedure of using relax. In
> modelfree, we usually fit data into three models e.g. isotropic, axially
> symetric or totally anistropic, and in each model for example isotropic, the
> procedure will be 
> 1. estimate tm from T1/T2 or other programs
> 2 fit each residue into five models
> 3 select the best model for each residue
> 4  Fitting and simulating tm and all model free parameters simmutanueously 
> using
> the selected model for each residue (with error analysis)
>
> What about relax? Do I need to run 4? I am trying to Minimise all parameters
> first and then Monte Carlo Simulations, but it seems very slow.
>
> Cheers!
>
> Hongyan
>
> Quoting Chris MacRaild <c.a.macraild@xxxxxxxxxxx>:
>
> > On Wed, 2006-12-27 at 16:49 +0800, Hongyan Li wrote:
> > > Dear relax users,
> > > I have tried to run relax with my dynamic data. Using the simplest
> > isotropic
> > > model, I haved run mf-multimodel.py (without Monte Carlo simulations) 
> > > and
> > > modsel.py to select a best model for each residue. I would like to use
> > selected
> > > model to run again with Monte Carlo simulations like what I did in
> > Modelfree. 
> >
> > The simplest way to do this is probably to insert the Monte Carlo
> > simulations into the modsel.py script, immediately after doing the model
> > selection. So the last few lines of the script should look like:
> >
> > ...
> >
> > # Model selection.
> > run.create('aic', 'mf')
> > model_selection('AIC', 'aic')
> >
> > # Monte Carlo Simulations
> > monte_carlo.setup('aic', number=100)
> > monte_carlo.create_data('aic')
> > monte_carlo.initial_values('aic')
> > minimise('newton', run='aic')
> > eliminate(run='aic')
> > monte_carlo.error_analysis('aic')
> >
> > # Write the results.
> > state.save('save', force=1)
> > results.write(run='aic', file='results', force=1)
> >
> > > I
> > > wonder if there is a script for this purpose and how to float tm value
> > which
> > > was estimated accoring from T1/T2ratio, so that relax can also simulate
> > it.
> > >
> >
> > Again, this can be done by simple modification of the end of the
> > modsel.py script. Something like:
> >
> > ...
> >
> > # Model selection.
> > run.create('aic', 'mf')
> > model_selection('AIC', 'aic')
> >
> > # Minimise all parameters.
> > fix('aic', 'all', fixed=0)
> > minimise('newton', run='aic')
> >
> > # Write the results.
> > state.save('save', force=1)
> > results.write(run='aic', file='results', force=1)
> >
> >
> >
> > Note that because of the dimensionality of the function being minimised
> > here, grid search is not possible. Minimisation is likely to find only a
> > very local minimum. It is therefore important to do this only after
> > optimising dynamic parameters with respect to a good estimate of tm.
> >
> > It is good practice to iterate the whole proceedure until the result
> > converges.
> >
> >
> > Chris
> >
> >
> > > Any suggestion would be highly appreciated!
> > >
> > > Cheers!
> > >
> > > Hongyan
> > >
> > > Dr. Hongyan Li
> > > Department of Chemistry
> > > The University of Hong Kong
> > > Pokfulam Road
> > > Hong Kong
> > >
> > >
> > > _______________________________________________
> > > relax (http://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
>
>
> Dr. Hongyan Li
> Department of Chemistry
> The University of Hong Kong
> Pokfulam Road
> Hong Kong