[gmx-users] g_rms questions

[Corianne Feenstra-Vos]

ChenYanjun wrote:
> Hi all,
> 3)Are intrinsic difference between Rho and RhoSc? How to choose? Can they
> deal with peptides short as 5 residues.Is there some problems if I apply
> g_rms to compare peptides such as loops or secondary structures which are
> not quite globular.

If you read the reference paper from the g_rms manpage carefully, you
should be able to understand the differences between Rho and RhoSc. It
has been too long for me since I last saw that article, so I can't help
you there. (You could always try to contact the authors.)

The Rho parameters were specifically designed to give comparable values
for similarity independent of the chain length, so you could for example
see whether two conformations of a 10 residue peptide are more similar
than two conformations of a 200 residue protein. With RMSD, that won't
work since the protein will have a larger rmsd by default (has more atoms).

There is in principle, I believe, no restriction for the Rho parameters
to be applied to globular structures only. I've played around a lot with
different conformations of peptides, calculated the Rho_sc (I believe
I used that one the most), and looked at the 'visual' differences of
the fitted structures with, e.g, rasmol. See if your intuitive idea
of similarity coincides with the numbers you get.

As far as I am aware, there is no lower limit to the applicability of
the Rho parameters with respect to sistem size, i.e. it should work for
anything consisting of more than 3 atoms.


Groetjes,

Anton