VMD-L Mailing List

From: Chris Neale (candrewn_at_gmail.com)
Date: Thu Jun 10 2021 - 12:26:39 CDT

Thank you all for the wonderful suggestions. I will certainly give them all
a try.

My actual use case is something like docking, where I want to view a
"trajectory" with hundreds or thousands of frames that all have the same
protein target but each frame has a different chemical bound, possibly in a
different location, and certainly with different numbers/types of atoms in
the small molecule component. I want them in the same molecule because
activate/deactivate commands (or GUI) for different molecules is more
cumbersome than pushing the frame slider along (though this makes me
realize that just having a slider that activates/deactivates separate
molecules in turn would do what I need after cloning a representation). Our
current solution is to script over auto-image generation with VMD and then
combine the output images into a movie. It's OK, but the interactive
component is really tiresome (identify a frame of interest, load that
separately into VMD, interactive visual inspection one at a time).

Since MGLtools looks like it's written to work with docking outputs, that
sounds promising and I'll give it a go today (thank you Raman). I'll also
try Pawel's suggestion to use Pymol (I've played with it in the past but
only for "normal" trajectories and never thought to check if it had the
functionality to do what I'm looking for). I'll take further discussions on
those softwares to their respective mailing lists, but I do really
appreciate the heads up here.

Josh's topotools code successfully wrote the .psf and .pdb files, which I
then read in and got almost what I am looking for except everything was put
into a single frame, so visual inspection going from one "frame" to the
next is cumbersome and involves selection by residue range. Josh: (1) thank
you for providing the code snippet, which made testing this out very fast.
Also, (2) does a topotools solution to this already exist?

Thank you,
Chris.

On Thu, Jun 10, 2021 at 8:19 AM Vermaas, Josh <vermaasj_at_msu.edu> wrote:

> Hi Chris,
>
>
>
> Can’t you use topotools mergemols to do this?
>
>
>
> package require topotools
>
> set pdblist [glob *pdb]
>
> set midlist [list ]
>
> foreach pdb $pdblist {
>
> set mid [mol new $pdb]
>
> lappend midlist $mid
>
> }
>
> set mol [::TopoTools::mergemols $midlist]
>
> animate write psf merged.psf $mol
>
> animate write pdb merged.pdb $mol
>
>
>
> This would get you everything into a single molecule, but as you state,
> this would be a royal mess to select over. What exactly are you trying to
> do that needs them in the same molecule?
>
>
>
> -Josh
>
>
>
> *From: *<owner-vmd-l_at_ks.uiuc.edu> on behalf of Chris Neale <
> candrewn_at_gmail.com>
> *Date: *Wednesday, June 9, 2021 at 8:32 PM
> *To: *vmd-L <vmd-l_at_ks.uiuc.edu>
> *Subject: *vmd-l: loading multiple different topologies into the same
> "molecule"
>
>
>
> Dear users:
>
>
>
> does anybody know if it's possible to load many different PDB files into
> the same molecule and have those PDB files have different system
> compositions? An example would be to be able to run "vmd -f *.pdb" on a
> large collection of PDB files of the same protein family, etc. I realize
> that some selection tools would then break (e.g., a selection of "resid 19"
> might not always play nice), but sensible selections, like "protein"
> "water", etc might be OK. I know that there are some good ideas with
> putting particles far away for things like constant pH simulations, but
> that's not as generalizable as loading in all SH2 domains in the PDB, for
> example.
>
>
>
> I realize that I could script image generation, or load hundreds of
> separate molecules with a "vmd -m *pdb" command, but none of those is as
> intuitive for browsing structural data where some models may have missing
> residues, insertions, different sequences, different numbers of water, etc.
>
>
>
> If such a capability does not exist, can anyone guess how hard it would be
> to implement right in the GUI? I'm not asking for anyone to actually do
> this, just trying to get a sense of how hard it would be.
>
>
>
> PS: I've got 10+ years experience with VMD, and I'm fairly confident that
> this is not possible at the present time. However, I'm asking since (a)
> somebody might have a great alternative solution and (b) in any event I'm
> interested in learning how hard this would be to implement.
>
>
>
> Thank you for your advice,
>
> Chris.
>