Re: Speed up of the TSMFK01 model. -- September 09, 2013 - 21:25

Hi Edward.

Where should I start write a test for the conversion?

Should it be a system or unit test ?

And should I load a dataset, minimize it, and then see
if the converted parameter is available for the model?

Best
Troels



Troels Emtekær Linnet


2013/9/9 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:
> Hi.
>
> > There is an explanation for everything, including user errors. :-)
>
> I'd call this more developer errors ;)  Talking about development, I
> will soon ask the other relax developers about you joining the
> development team.  This will have to wait a little while as I am on
> holidays for about 3 weeks starting from this weekend.
>
>
> > I have analysed the 0.48 M GuaHCl dataset (the folded protein), and so
> > this is not comparable to the
> > 1M GuaHcl (intermediate between folded/unfolded) dataset, which is
> > shown in the figures in the paper.
> >
> > So, I got the original datapoints for figure 3, which shows the
> > ln(k_a) per GuaHCl.
> > And now k_a fits perfect for 0.48 M.
> >
> > And I have today analysed the 1M dataset.
> >
> > Everything matches until first digit, so I am satisfied.
>
> This sounds great!  The model is so incredibly simple that finding out
> that a data mix up was the source of the problem is quite a relief.
>
>
> > So, I will soon send a swarm of patches to include this dataset.
>
> All patched 
> (http://thread.gmane.org/gmane.science.nmr.relax.devel/4372/focus=4563).
>
>
> > Thanks for looking!
> >
> > I will look into the tsmfk01 code to speed it up.
>
> No problems.  I look forward to seeing your solutions.
>
>
> > To compare to the numerical methods, you mentioned that one could make
> > auto conversion of the
> > parameters?
> >
> > So could k_AB be calculated for the numerical methods, so:
> > k_AB = kex*(1-pA)
>
> Yes, this can be done.  The complication is handling the Monte Carlo
> simulations, specifically the parameter error estimate calculation at
> the end of all simulations.  You can have a go at this yourself, if
> you wish.  The key is to look for where the 'tex' or 'pB' parameters
> are handled, as these are not parameters in most of the dispersion
> models, and to study this code.  These two parameters will be handled
> as you wish k_AB to be.  If k_AB is calculated, then so should k_BA.
> Maybe phi_ex should also be added to this list of 'auxiliary'
> parameters so that the numeric models can be easily compared to the
> fast exchange models.  Are there any other auxiliary parameters you
> can think of?  What about the alpha factor which determines if the
> exchange is fast, slow or intermediate?
>
> Cheers,
>
> Edward