[gmx-users] Discrete structure factor from g_rdf

[comcon1 at erg.biophys.msu.ru]

> On 02.06.2010, at 09:37, comcon1 at erg.biophys.msu.ru wrote:
>
>> I'm trying to repeat some X-ray diffraction analysis on lipid bilayers
>> according to the article:
>>
>> http://dx.doi.org/10.1529/biophysj.104.046821 (it's free)
>>
>> I have several corrected OPLS all-atom bilayer models and I need to
>> validate my model according to experimental x-ray diffraction data. But
>> I
>> can't understand, how '-sq' property of g_rdf can helps me to compute
>> structure factor.. May be anybody have some experience in this theme?
>>
>
>  So far I also haven't really used g_rdf to calculate the structure factor
> of a system, so I can't give you a hint how to reproduce the results in
> the paper. However as the version of g_rdf in the current and earlier
> release was not able to treat every atomtype perhaps you should at first
> consider the git version of g_rdf for your calculations. This version is
> capable of treating every atomtype if you supply the corresponding data.
> The calculation is based on a so called Cromer-Mann fit that requires
> different parameters for different atoms. This parameters can be for
> example obtained from
>
> http://www.ruppweb.org/xray/comp/scatfac.htm
>
> There is a reference to an article and a short introduction contained on
> the website.
>
> Finally append your parameters to the list in share/top/sfactor.dat and
> use the corresponding flag in the command line ( g_rdf -h will tell you
> which)
>
> /Flo

Thx for ans!

I have already done this. I also modify sfactor.c a little to select atoms
acording to atom number instead of atom name ;-) But `g_rdf -sd` gives me
plot of 'absorbtion units' versus 'reverse nanometers'. And it looks
absolutely different from structure factor from article I cited. For
example in article SF is fading sinusoid while g_rdf gives me positive
increasing curve. I suppose it's diffraction spectrum, not structure
factor. Or ... ?

(i rly have no idea, so i'm scanning source code of sfactor.c now, but may
be anybody know what `g_rdf -sd` exactly do?? )

Thanx before,
comcon1)