[gmx-users] g_rdf and number of atoms to include

[Omer Markovitch]

On Wed, Oct 21, 2009 at 07:28, Enemark Soeren <chees at nus.edu.sg> wrote:

>  Dear users,
>
> I would like to compare interactions between molecules by using RDF. I have
> tried looking at glycine and water, and compare the following two
> interactions:
>
> 1)      between the amine hydrogen atoms in glycine and the oxygen atom in
> water
>
> 2)      between the carboxyl oxygen atoms in glycine and the oxygen atom
> in water
>
> However, my result in 1) depends on how many of the 3 hydrogen atoms I
> include in the calculations. Why is that?
>
If you mean that when focusing your RDF calculations on either one of the
three hydrogens results in three different RDFs then it means that each
hydrogen feels water differently. I bet this difference is only for the
first peak of g(r) and the other peaks overlap between the three RDFs. As
for how reasonable this result, its not unlikely because glycine has atleast
2 different types of hydrogens (say, C-H vs. N-H), depending on the
protonation state. You might want to provide more details on the system you
are studying to get better answer.
I am assuming that the RDFs are converged so that including more trajectory
data and/or sampled molecules and/or changing bin size does not result in a
drastic change to the curves.


> Does that mean that I cannot directly compare the strengths (RDF peak
> height) of the two interactions as they are not based on the same number of
> atoms? Does it also mean that I must always calculate RDFs by using 1 atom
> on each of the particles/groups that I am comparing?
>
I am not sure how good it is to use the first peak of g(r) to analyze
strengths, but you should also consider the width and area under peak. This
peak is an average on all nearest neighbours, bonded or not, so it might not
give you a good estimate of the hydrogen bond, for example.
If you are unsure of your g(r) calc it just for water (that is - only
oxygen-oxygen of water-water). At long distances (~10 Angstroms) it should
fluctuate around 1.

Bests, Omer Markovitch.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20091021/16a81837/attachment.html>