Re: vmd-l: error using Namdenergy

From: Chandra Ramananjara (
Date: Wed Jun 21 2006 - 21:17:13 CDT

Hi Peter,
Thanks a lot for the tcl script. It worked even on the PDB files with the CRYST header.


Peter Freddolino <> wrote: Hi Chandra,
I can reproduce the crash with your file. It looks like an odd
interaction between the dcd writers of vmd and namd actually. I'll have
a look at it with the devs tomorrow. In the meantime, I'm attaching a
patched version of namdenergy that should work around this problem;
please just source it before you open the namdenergy gui.

Chandra Ramananjara wrote:
> Hi Peter,
> Thanks for your email. The version of VMD that I'm using is VMD for
> Linux, version 1.8.4b6 (December 22, 2005).
> Thanks, attached to this email is the psf/pdb combo in which the pdb
> file has the CRYST header, I tried namdenergy on this combo just now
> and I'm still getting the null results. The pdb file was created using
> the writepdb command (in VMD) to save the coordinates of a particular
> frame (frame 78) of the trajectory.
> Yes you're right, if namdenergy is run on a loaded dcd file, it will
> loop over the trajectory, and the selected atoms of interest (whose
> interaction energies are to be calcuated) will be updated in each
> frame. My case is different. In each frame, the atoms of interest
> (whose interaction energies I want to calculate) are different from
> frame to frame so its not as straightforward as using just one
> selection of atoms for all frames. I think the example script I posted
> in my last email must have confused you, because the atoms of interest
> were all oxygens. I just did this to simplify the script.
> In the actual script I'm writing, the selection of the atoms of
> interest (whose interaction energy I want to calculate) is more
> complicated and the atoms of interest vary from frame to frame so
> using one selection text (e.g. [atomselect top "name C"]) for all
> frames in the trajectory is not an option.
> Do you think it might be my version of VMD ? I think the version I'm
> using (1.8.4b6) is a relatively new version. The pdb/psf combo I'm
> having problems with are attached to the email. Thanks.
> -Chandra
> */Peter Freddolino
/* wrote:
> Hi Chandra,
> a couple things here... could you let us know what version of VMD
> you're
> using, and on what platform?
> Namdenergy uses NAMD itself for the actual energy calculations so it
> seems very odd that changing the CRYST header causes a problem
> (indeed,
> I have used it many times in cases where it hasn't caused a problem,
> hence my confusion--in fact, it just worked ok for me now).
> Perhaps you
> could send me one of the pdb/psf combos that fails?
> Also, just for your information, if you run namdenergy on a loaded
> dcd
> it will automatically loop over the trajectory; there's no need to
> run
> each frame separately.
> Peter
> Chandra Ramananjara wrote:
> > Hi all,
> > I am observing a strange error in namdenergy. I am using the
> > text-based version of namdenergy in VMD (in text mode) which can be
> > called using the command 'package require namdenergy'.
> > Here is the problem. I have a dcd file and I want to step
> through each
> > frame and calculate energies for interactions between various
> atoms.
> > So, for each frame of the dcd file, I select all the atoms and
> write
> > them to a pdb file (called temp.pdb) and then load the pdb file
> (along
> > with a general psf file) so I can run namdenergy on the atoms in
> that
> > PDB file. So, for example, the script for looping through 1000 dcd
> > frames is something like this:
> >
> > for { set i 0 } { $i < 1000 } { incr i 1 } {
> >
> > molinfo top set frame $i
> > set all [atomselect top all]
> > $all writepdb TEMP.pdb
> > mol load psf base.psf pdb TEMP.pdb
> > set group [atomselect top "name O"]
> > namdenergy -nonb -sel $group -par param.inp
> >
> > }
> >
> > The problem I encounter is that the results I get from
> namdenergy are
> > always 0 (obviously wrong).
> > But when I open the TEMP.pdb file (generated in VMD using
> writepdb) I
> > notice the first line in the PDB file is something like: "CRYST1
> 24 24
> > 24 90.0 90.0 90.0". This line is the first line in every PDB file
> > generated using VMD's writepdb command. It specifies the unit cell
> > length and angles.
> > I notice that if I delete this line, reload the PDB file into
> VMD and
> > re-run namdenergy, I get the right answer for the energy. It
> seems the
> > namdenergy program cannot recognize PDB files with the "CRYST1...."
> > header.
> > Unfortunately, all PDB files generated using the writepdb
> command in
> > VMD have this header. As I mentioned, manually deleting this header
> > allows namdenergy to recognize the PDB file and gives right
> answers.
> > But since I want everything to be automated, programmatically
> (using
> > Tcl file commands) deleting the header of every generated PDB file
> > would use up valuable computational time. Is there any workaround ?
> > And does anybody have any idea why namdenergy refuses to
> recognize PDB
> > files with the "CRYST1...." header ?
> >
> > Thanks,
> > Chandra
> >
> >
> ------------------------------------------------------------------------
> > Do you Yahoo!?
> > Get on board. You're invited
> >
> > to try the new Yahoo! Mail Beta.
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around

Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:42:15 CST