[gmx-users] Problem with g_rdf

[Justin Lemkul]

On 2/19/13 5:04 PM, Thomas Schlesier wrote:
> On 2/19/13 3:10 PM, Thomas Schlesier wrote:
>  > Dear all,
>  > I have a question regarding g_rdf.
>  > My system consits of 10 disacchardids (in the following A-B, with
>  > A-ring/monosaccharid and B-ring/monosaccharid) and i want to determine the RDF
>  > (with flag '-rdf mol_com')for A-A, B-B and the intermolekular for A-B
>  >
>  > A-A and B-B is no problem.
>  > But for A-B g_rdf determines the intermolecular contribution (which i want) AND
>  > the intramolekular contribution (which i wanted to be excluded).
>  >
>  > Any (nice) ideas to solve this problem for GMX 4.0.7?
>  >
>  > Long way would be to calculate each intermolecular contribution individual and
>  > average over all. Would require some scripting :( so any better solutions are
>  > welcome.
>  >
>
> Create a topology with a value of nrexcl sufficient to exclude all
> intramolecular interactions and create a .tpr file from it.  Use that as input
> to g_rdf.
>
> -Justin
>
> -------------------------------------------------------------------------------------
>
>
>
> Thanks for the suggestion. But it doesn't really fix the problem (or i don't get
> it).
> Ok what i have done (as an example only 3 molecules, A1 are all atoms of ring A
> of molecule 1, other stuff analogous)
>
> Index group:
> [ A ]
> A1  A2  A3
> [ B ]
> B1  B2  B3
> [ both ]
> A1  B1  A2  B2  A3  B3
>
> RDF(A-A): chose two times group A
> RDF(B-B): chose two times group B
> RDF(A-B): ?
>
> If i chose group A and B i get the same result with 'nrexcel 3 or 10', i.e.
> mixing of intra- and intermolecluar A-B distances.
> If i chose 'both' two-times, i get the impression the the result is the RDF
> between the whole disaccharides, i.e. A1 and B1 are recognised as one molecule
> (AB)1 and i get the RDF between:
> (AB)1 - (AB)2 ; (AB)1-(AB)3 ; (AB)2-(AB)3
>
>
> I tested this also with an index-file with only 1 molecule.
> RDF(A-B) gives the intramolecular RDF
> RDF(both) is zero, because the is only one molecule, center of mass distance
> must be zero.
> This also doesn't depend on using 'nrexcel 3 or 10'
>
>
> Did i made a mistake? Or any further ideas?
>

No, I missed a piece of information in your original post.  The nrexcl approach 
only works with -rdf atom.  I would suspect that if you want to use -rdf 
mol_com, you will have to have individual groups for which you want the RDF 
around the central group.  You can analyze all the B groups around each A (and 
vice versa, if you need to) using the -ng flag.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================