[gmx-users] On creating pdb file using dynamic indices from g_select

[Mark Abraham]

On 2/09/2012 6:03 AM, Mr Bernard Ramos wrote:
> Hi everyone!
>
> I intend to create a trajectory file (in pdb format) of atoms dynamically indexed by g_select. However, I failed to produce the correct pdb file. Here is how I did it. Kindle help please.
>
> 1. I created the index file using g_select for water oxygen atoms within 0.4 nm of Atom 3
>
>
> g_select -f A.xtc -s A.tpr -select 'resname SOL name OW and within 0.4 of atomnr 3' -on X_ndx.
>
> The first few lines of X.ndx look like this
>
> [ resname_SOL_and_name_OW_and_within_0.398_of_atomnr_3_0.000 ]
>
> [ resname_SOL_and_name_OW_and_within_0.398_of_atomnr_3_0.200 ]
>   685
> [ resname_SOL_and_name_OW_and_within_0.398_of_atomnr_3_0.400 ]
>   685 1045
>
>
>
> 2. Then I went into trjconv to create a pdb of these water oxygens
>
> trjconv -f A.xtc -s A.tpr -o X.pdb -n X.ndx
>
> After keying in Enter, I see a list of the the dynamic groups that looks like the ones above. I have 10000 groups in the list corresponding to 10000 frames but within each frame having different number/indices of water oxygens. I then get a prompt asking me to "select a group". I'm not quite sure what number to type but I tried entering "0", "0 to 10000", but both gave me wrong pdb files. They dont match with the indices given in X.ndx. How do I tell Gromacs to create a pdb file for these specific atoms? And next time I use a dynamic index, what do I type Gromacs asks to "select a group"? Thanks..

No GROMACS tool can interpret this index file as a dynamic selection, 
because the tool cannot map the time of a trajectory frame to the 
selection in the index group. It is a feature expected in GROMACS 5.0, 
but don't hold your breath. In the meantime, you will have to write a 
shell script to loop over frames, applying the correct single index 
group to each frame. trjconv -sep might be a good starting point.

Mark