[gmx-users] Fitting when structure and reference are slightly different

[Ran Friedman]

Hi Jochen,

If you're dealing with protein structures you can use multiprot:
http://bioinfo3d.cs.tau.ac.il/MultiProt/

I put a runner for gromacs (do_multiprot) on the user contribution page
some time ago, and can put up a more recent version if it doesn't work
with GMX 3.3.3.

Ran.

Jochen Hub wrote:
> Hi,
>
> I run frequently into the problem that I want to fit a structure on a
> reference which is, however, not completely identical to the
> structure. The reference could, e.g., have a different protonation
> state or have an additional residue at a terminal. Is such a case a
> backbone-fit using trjconv gets erroneous since the atomic numbers of
> the backbone atoms in the two structures are not the same.
>
> The best solution would be the possiblity to give trjconv 2 index
> files, one giving the N fit atoms of the reference and the other the N
> related atoms (but with different atom numbers) of the fitted
> structure. As far as I know, something like that is not supported by
> gromacs so far, right?
>
> How do you usually deal with such a problem? Is there some more or
> less easy workaround?
>
> By the way, removing the atoms that do not agree in the two structures
> is not an option since I want to start a simulation from the fitted
> structure.
>
> Thanks a lot,
> Jochen
>