Re: Using Model-free models

[Mahdi, Sam]

Hi Edward,

So there is no point in running model-free unless you have data at 2 field
strengths? I also had a slight side question. I've read your papers on the
diffusion seeded paradigm, and dynamic models 1 and 2. I understand the
importance of the diffusion tensor value, and understand how the d'Auvergne
protocol fixes that. My question is, does the diffusion tensor help the
model selection then? I've run the new protocol on a protein that had data
at 2 different frequencies. I then found out what model it picked for each
residue. Then I ran the model free models with the estimated diffusion
tensor, and selected the model I wanted it to run, to see whether or not I
would obtain a similar value to the d'Auvergne protocol run for those
residues, which it did (i.e. for residues that had S^2 from m4 in
d'Auvergne protocol, the S^2 from my m4 single model free run gave almost
the same values for those residues). From that I assumed, model free models
data would give similar values to the d'Auvergne protocol, if the correct
model was selected; however, I wasn't able to get any model selection which
is why I sent the initial email. That's why my question is, is the
diffusion tensor crucial then for selecting the proper model?

Sincerely,
Sam

On Fri, Nov 11, 2016 at 1:21 AM, Edward d'Auvergne 
wrote:

> On 10 November 2016 at 23:50, Mahdi, Sam 
> wrote:
> > Hello everyone,
> >
> > I've been attempting to run model free models (due to only have data at
> one
> > field strength), but I've come across a problem. I have set up both the
> > model free models script and the model free model selection, both models
> > run fine, and I do end up receiving data that does make sense and is
> good;
> > however, the script doesn't seem to actually select any models. I.e. if I
> > put in the m0-m9 in the script, it'll select m9 for every single reside
> and
> > only give me Rex data and no S^2. If I run m0-m4 same thing, it'll run
> > every residue at m4, instead of running through all the models and
> > selecting the data that fits that model. I know it states that the model
> is
> > supposed to loop through every model, then you run the model selection to
> > select what model fits which data set, but I don't know if my script is
> set
> > up incorrectly or what. Here is my script for model-free models:
> > pipes = ['m0','m1','m2','m3','m4']
> > for name in pipes:
> > pipe.create(name,'mf')
> > sequence.read('ssNOE_RGS4Agnes',res_num_col=1)
> > spin.name('N')
> > spin.element(element='N',spin_id='@N')
> > spin.isotope('15N',spin_id='@N')
> > structure.read_pdb('cluster1_12.pdb',read_mol=1)
> > relax_data.read(ri_id='R1_Agnes',ri_type='R1',frq=599.642*1e6,
> > file='R1_RGS4Agnes',res_num_col=1,spin_name_col=2,data_
> col=3,error_col=4)
> > relax_data.read(ri_id='R2_Agnes',ri_type='R2',frq=599.642*1e6,
> > file='R2_RGS4Agnes',res_num_col=1,spin_name_col=2,data_
> col=3,error_col=4)
> > relax_data.read(ri_id='ssNOE_Agnes',ri_type='NOE',frq=599.642*1e6,
> > file='ssNOE_RGS4Agnes',res_num_col=1,spin_name_col=2,
> data_col=3,error_col=4)
> > diffusion_tensor.init(1e-8,fixed=True)
> > sequence.attach_protons()
> > interatom.define(spin_id1='@N',spin_id2='@H',direct_bond=True)
> > interatom.set_dist(spin_id1='@N',spin_id2='@H',ave_dist=1.02*1e-10)
> > structure.get_pos('@N')
> > structure.get_pos('@H')
> > interatom.unit_vectors()
> > value.set(-172*1e-6,'csa',spin_id='@N')
> > model_free.select_model(model=name)
> > minimise.grid_search(inc=11)
> > minimise.execute('newton')
> > results.write(file='results',force=True)
> > state.save('save',force=True)
> >
> > Again I only have it at m0-m4 because I wanted to obtain the m4 data, but
> > even if it is to m9, it'll only run the m9 model (i.e. I will only
> receive
> > Rex data because that's whats attributed to m9).  Here is my model
> > selection script:
> > pipes = ['m0','m1','m2','m3','m4']
> > for name in pipes:
> > print("\n\n#"+name+"#")
> > pipe.create(name,'mf')
> > results.read(file='results.bz2',dir=name)
> > eliminate()
> > model_selection(method='AIC',modsel_pipe='aic')
> > state.save('save',force=True)
> > results.write(file='results',force=True)
> >
> > Again, this is only set up to m4 for the same reasons as before. I don't
> > know whether or not I have something in my script set up, but my final
> > results file just shows the model that was selected for every residue, is
> > the last model I have in my pipes (i.e. if m4 is the last model, it'll
> run
> > model 4, if m9 is the last one, it'll run m9). In this sence it appears
> to
> > be a single model scenario even though I've set up to be for multiple
> > models. Thank you again in advance.
>
> Hi Sam,
>
> This is perfectly normal!  The key is the line:
>
> diffusion_tensor.init(1e-8,fixed=True)
>
> Here you have fixed the diffusion tensor to spherical tumbling of 10
> ns.  Note that >80% of the contribution to relaxation is due to the
> global tumbling. 

Re: Dispersion Back Calculation

[Edward d'Auvergne]

On 27 October 2016 at 18:10, Jeremy Anderson  wrote:
> Hi Edward and Troels,
>
> Thanks for pointing me in the right direction.  I had dug around a bit in
> the test_suite directory but wanted to make sure I was looking in the right
> place before I descended into the rabbit hole.
>
> I got the back calculation to work using the
> ./test_suite/shared_data/dispersion/ns_mmq_3site_linear/relax_results/solution.py
> script pretty much as-is, just changing the spin parameters to my liking,
> calculating the curve, and outputting the values (ignoring the data and
> residuals in the output file).
>
> Something I didn't mention is that the reason I've been importing the models
> into ipython is so I can hold parameters constant through my own grid search
> and minimization functions, which I had found somewhere in the documentation
> was not possible inside relax for the minimization.  I originally thought
> this would be easier outside of relax.
>
> The reason for this is because I'm in a situation where I can observe HSQC
> peaks in slow exchange in one variant and skewed populations of one or the
> other peaks in two other variants.  I've been working on using the
> complementary information, in this case the observed dw and the kex from ZZ
> exchange experiments, to investigate multi-state exchange in all variants.
>
> The chem. shift differences of the two skewed variants match the measured
> nicely but the rates from CPMG are ~20 fold higher.  Therefore I wanted to
> check and see if a 3-state model with some parameters held constant would
> have infinite solutions (my assumption) or pop out something interesting and
> be able to distinguish between a couple models of the conformational process
> that I have in mind, which seems like a long shot.
>
> Sorry if thats too much information/way too open-ended but I figured I would
> give some context to the greater situation I have found myself in.  Thanks
> again!

Hi Jeremy,

It is true that you cannot fix a parameter in relax and optimise the
others.  The reason is two-fold.  Firstly the minfx library (
https://gna.org/projects/minfx/ ) does support this functionality.
Secondly, this functionality would be highly abused and a lot of
rubbish results will appear in the scientific literature, with a
detrimental effect on the reputation of the whole NMR field.

Also, I didn't think it was worth the time investment compared to
expanding relax to handle multiple data types at the same time, and
then optimising one set of parameters for all experimental data
simultaneously.  In your case, that would be loading the ZZ exchange
and CPMG data at the same time, and optimising the single model.  This
would be interesting, as the two experiment types contain both
complementary and overlapping information content.  So saying that the
overlapping content should only come from the ZZ experiment might
over-constrain the CPMG experiment due to any biases or experimental
noise from that experiment.  Are you able to set up the problem in
this alternative way in iPython?

Regards,

Edward

___
relax (http://www.nmr-relax.com)

This is the relax-users mailing list
relax-users@gna.org

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: relax-users Digest, Vol 116, Issue 38

[Edward d'Auvergne]

On 29 October 2016 at 23:44, Mahdi, Sam  wrote:
> Hi Edward,
>
> I was reading the theory on model free within the manual and I had a quick
> question. The d'Auvergne protocol, is that the model-free analysis in
> reverse with the universal solution?

Yes, that is the technique.

> Or is that the model-free models with
> only the AIC model selection (no universal solution). Both methods were
> under the new-protocol section, so I was confused a bit as to which one the
> d'Auvergne protocol is running.

This is a fragment of the protocol.  It is one part of the iterative
procedure shared with the methods that require an initial, external
diffusion tensor estimate.  The key references to understand all of
this are:

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. (http://dx.doi.org/10.1039/b702202f)

d'Auvergne, E. J. and Gooley, P. R. (2008). Optimisation of NMR
dynamic models II. A new methodology for the dual optimisation of the
model-free parameters and the Brownian rotational diffusion tensor. J.
Biomol. NMR, 40(2), 121-133.
(http://dx.doi.org/10.1007/s10858-007-9213-3)

Regards,

Edward

___
relax (http://www.nmr-relax.com)

This is the relax-users mailing list
relax-users@gna.org

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: Using modified all models free

[Edward d'Auvergne]

On 1 November 2016 at 22:12, Mahdi, Sam  wrote:
> Hello everyone,
>
> I had a quick question regarding running the single model free run (m4).
> For some proteins we only have data at only one field strength, so I know I
> can't use the d'Auvergene protocol (needs a minimum of 4), so I was
> thinking I could run the all models free model. The problem is, I know
> models m6-m9 are models for another field strength, but since I don't have
> data at another field strength, could I just remove them? Or will relax
> just not run m6-9 since I've only loaded one field strength?

Hi Sam,

The protocol I designed for iteratively optimising and fine tuning the
coupled diffusion tensor + model-free optimisation and model selection
problem will only work with 2 or more fields.  If you would like to
understand this problem, please read:

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. (http://dx.doi.org/10.1039/b702202f)

From this, hopefully you will understand why the NMR field will not
take any results from a single field strength analysis seriously.
Note that this has been the case since the mid 1990's.

Regards,

Edward

___
relax (http://www.nmr-relax.com)

This is the relax-users mailing list
relax-users@gna.org

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