[MMTK] Questions on nMoldyn trajectory conversions and Sqw units

Konrad Hinsen research at khinsen.fastmail.net
Fri Aug 5 14:29:30 UTC 2016

Dear Katy,

I don't have much experience with GROMACS simulations, but I can at 
least provide some background information that might help you.

> 1.       I would like to use nMoldyn to analyse trajectories output by
> GROMACS. I have converted the GROMACS trajectory output to .dcd format
> using Wordom, and have tried using the NAMD and CHARMM conversion
> options in nMoldyn to convert to the appropriate MMTK format. The
> conversions claim to be successful, but when I load the generated .nc
> file it seems that the periodic boundary conditions of my system have
> been lost (nMoldyn recognises an infinite universe).

There are different variants of the DCD that differ in particular in how 
they manage periodic boundary conditions. I suppose that in your 
conversion chain, different programs make different assumptions about 
which dialect of DCD to use. nMOLDYN distinguishes CHARMM and NAMD for 
that kind of reason, though I cannot explain the exact differences 
without looking at the code again - my brain isn't good at memorizing 
that kind of messy details.

> Does anyone have experience using nMOLDYN with GROMACS? If so, could you
> please let me know how you converted the relevant files to MMTK format?

I have a little-tested direct converter from GROMACS XTC + PDB to MMTK's 
netCDF files. All I can say for now is that it works for two 
trajectories I have (but which were produced by someone else), whereas 
it doesn't work for the only other person who tested it. Since he can't 
send my his trajectories, I am a bit stuck with this.

If anyone knows of some published GROMACS trajectories that anyone can 
use and copy without restrictions, please let me know. I do not know 
enough about GROMACS to produce reasonable trajectories myself.

> 2.       The documentation on the dynamic structure factors (coherent
> and incoherent, DCSF and DISF) in the nMoldyn User’s Guide suggests that
> the output should have units. However, when I use ncdump on the
> generated .nc  file, the Sqw ouput is listed as unitless. Could someone
> confirm exactly what form (and units) of the dynamic structure factors
> are being output by the DCSF and DISF functions in nMoldyn?

I can check with the people who wrote the code in about a month (it's 
holiday time in France right now). I'd expect structure factors to be 
dimensionless by definition (i.e. by virtue of a suitable normalization 
factor), but I am not familiar with the conventions used in neutron 
scattering, so I'd better shut up.

Konrad Hinsen
Centre de Biophysique Moléculaire, CNRS Orléans
Synchrotron Soleil - Division Expériences
Saint Aubin - BP 48
91192 Gif sur Yvette Cedex, France
Tel. +33-1 69 35 97 15
E-Mail: research AT khinsen DOT fastmail DOT net
ORCID: http://orcid.org/0000-0003-0330-9428
Twitter: @khinsen

More information about the mmtk mailing list