[gmx-users] Another shuffling question ..

[David van der Spoel]

On Wed, 1 May 2002, Marc Baaden wrote:

>To give a practical example:
>I have the original system, which I use to create tpr files and eg index
>files needed for analysis.
>
>Now the original run is shuffled + sorted and so indices in the index
>files and in the trajectory will not correspond.
>
>With "deshuffling" I can reverse the shuffling, but not the sorting, so
>now my indeces should correspond to the right moleculetype (eg protein
>is protein, water is water, ions are ions), but if I want to address eg
>a specific water molecule from my index file, it might have another
>number due to sorting, so you would have to recreate the index files
>from the deshuffled trajectory ?
>
>Is this correct? Does it make sense ? Or am I talking rubbish ?
Yes this is correct. Sorry for the mess, but the sorting was just a quick
hack to speed up things. Since it worked so well for some systems we left
it in.

>The overall message would then be: you have to create your index files
>from the final trajectory that you want to analyze, once it was simulated
>and deshuffled, and not from the starting point.
Most analysis program don't care about the order of atoms in your traj, as
long as  you have a good index. You can make an index from your
shuffled and sorted tpr file, so that you don't need to deshuffle at all.
In some cases (e.g. following that special water molecule) this may be a
problem...

Groeten, David.
________________________________________________________________________
Dr. David van der Spoel, 	Biomedical center, Dept. of Biochemistry
Husargatan 3, Box 576,  	75123 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://zorn.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++