Re: [h5md-user] H5MD for proteins

[Pierre de Buyl]

Hi Konrad,

On Mon, Sep 09, 2013 at 12:15:01PM +0200, Konrad Hinsen wrote:
> I have been rather quiet on this list for a while, mostly because I
> was busy with other things, but I still read the messages to follow
> progress, which looks very encouraging.
> 
> As I mentioned earlier (long ago by now), for the question of
> trajectory storage is inseparable from the question of storing a
> definition of the system that is being simulated. There is no point
> (for me at least) in having a trajectory from which I can pick the
> coordinates of atom 532 without having the slightest idea of what atom
> 532 represents in my system.

H5MD attempts to be rather generic. The module/convention in the other thread
could be a satisfactory solution.

> I have been working on that aspect, which is not limited to
> trajectories, and came up with a data model for molecular simulations:
> 
>   http://bitbucket.org/molsim/mosaic/wiki/Home
> 
> Its HDF5 implementation should be compatible with H5MD, though I will
> be sure only once I try to use both together. That's my project for
> the near future.
> 
> A Mosaic universe defines molecular structures, a topology for the
> entire system (i.e. periodic boundary conditions and symmetry
> transforms), and how many copies of each molecule are present in the
> system. It provides an atom order that can serve as the link to the
> data in H5MD files. In HDF5, a universe is stored in a group, which
> could sit nicely next to the h5md group.

Do you mean that Mosaic-HDF5 would sit on top of H5MD?
 
> I'll post a few more specific questions about the current H5MD
> definition in separate threads, this message is just to provide
> the context for them.

Welcome back then :-)