[MMTK] Fitting data?

Konrad Hinsen hinsen at cnrs-orleans.fr
Tue Mar 23 17:53:22 CET 2004

On 23.03.2004, at 16:46, Jon Wright wrote:

> looks very impressive (thanks!). I'm wondering if it would be possible  
> to use it as a basis for fitting some powder diffraction data from  
> crystalline protein samples... The idea is to add an energy term  
> proportional to a chi^2

It is certainly possible, the only question is how much additional  
effort is needed.

>  function which describes the fit to the data. I think I can  
> (eventually) supply the value of this term, and it's derivatives with  
> respect to x,y,z co-ordinates

That looks like the main addition.

> infinite universe). There is a slight hitch, that for crystallography  
> the structure is described with the atoms smeared out by a "thermal  
> factor", but given the poor quality of powder data this can be just  
> left as a constant and

It's been ages since I looked at structure refinement (and even then it  
was only superficial), so please forgive me any naive questions...  
Isn't that just an additional set of parameters in the fit function?  
That would "only" take a modified minimization routine. Since that  
could be written in Python without a significant performance impact, it  
is perhaps worth trying.

> question. The manual (p13) mentions a developers guide, which I did  
> not manage to find yet - would someone direct me?

I guess I should remove that hint, as the developer guide still doesn't  
exist. In fact, I don't think I will ever write it as a separate  
document, rather as a set of contributions to the MMTK Wiki.

> From browsing the source, it looks like it should be possible to get  
> started with the example MMTK_harmonic_oscillator.c etc, with some  
> bits and

Yes. I am working on a much nicer way to implement force field terms,  
using Pyrex (http://www.cosc.canterbury.ac.nz/~greg/python/Pyrex/), but  
the current Pyrex version has a bug that makes this impossible, so I  
have to wait for a fix and I have no idea how long that will take.

> constructor in the associated .py file? I wanted to contact the list  
> before diving in, things like unit cell and space group probably  
> already have proper homes in other parts of the library, but the  
> treatment of symmetry

No, there is nothing like that yet. It should be fairly easy to add for  
someone who understands all the details (I don't at the moment). It's  
something I have wanted to do anyway, because I expect to work with  
crystallographic data in the future.

> stereochemical configurations which happen to fit our data as well.  
> Ideally I need some advice from someone who knows more about molecular  
> dynamics, because my background is mainly limited to powder  
> diffraction.

I can help out with MD stuff. Or I can put you in contact with protein  
MD people within 500 m of where you are.

> PS: If only you could understand how happy I was to find out that I  
> can "fix" the stereochemistry of a model by running it through a  
> minimiser with a little command line script! This is already a great  
> help if I can cycle models back

I do :-)  GUI tools have their place as well, of course, but in terms  
of flexibility nothing beats a Unix command line.

Konrad Hinsen                            | E-Mail:  
hinsen at cnrs-orleans.fr
Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-
Rue Charles Sadron                       | Fax:  +33-
45071 Orleans Cedex 2                    | Deutsch/Esperanto/English/
France                                   | Nederlands/Francais

More information about the mmtk mailing list