VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Tue Jun 25 2002 - 10:50:33 CDT

Hi,
  I'm forwarding a note from James Cavenaugh, as his note got bounced
originally:

----- Forwarded message from owner-vmd-l_at_ks.uiuc.edu -----

To: vmd-l_at_ks.uiuc.edu
From: sjcav_at_aros.net
Subject: rmsd functionality

Hello VMD folks:

        I've got some questions on the VMD RMSD functionality that I hope aren't
too basic. I'm running v. 1.7.2 on Windows 95. I have loaded 3 molecules,
the first being a time-averaged structure without coordinate fitting (using
AMBER's ptraj program) of a molecule of interest, the second the initial
starting molecule's coordinates, and the third being like the first but
with fitting turned on in the ptraj rms command. The third is the top
molecule. So I typed "vmdrmsd" at the VMD prompt and then toggled on the
top selection, and clicked the RMSD button, and after a while I got an
error message. Well I can't copy text with my mouse (using Windows 95,
left mouse button doesn't open a copy selection [as an aside, that would be
an option for improvement in future releases]), so I captured that in a
screen shot and am attaching it. Likewise for the RMSD window itself; it
is attached. I selected all the atoms. So, first question, what's going
on with these error messages? I did just a usual installation of VMD.

        I would expect that the rmsd vs. the three molecules (all are nbv.prmtop;
they were loaded with parm and crd option) should show numbers in the "RMS
Deviations" part of the window and not just at the bottom. Also, I would
expect that the number for the third nbv.prmtop in the list should be
0.0000, since that's the top molecule and is what I'm computing rmsd
against. But no numbers are shown except for a total rmsd at the bottom.
Why is this, and what exactly is this total rmsd?

        Clearly I can get an rmsd from ptraj, but it would be reassuring for my
understanding of both ptraj and VMD if I got the same numbers with both
programs. Here is the ptraj script used:

command: more ptraj.rms_av_vs_nbv
trajin NBV_av_0.85-1.2ns
reference ../NBV/Segment_0/nbv.inpcrd
rms reference mass out rms_av_0.85-1.2ns_vs_nbv.inpcrd_all nofit
rms reference mass out rms_av_0.85-1.2ns_vs_nbv.inpcrd_ncca @N,C,CA nofit
rms reference mass out rms_av_0.85-1.2ns_vs_nbv.inpcrd_all-fit
rms reference mass out rms_av_0.85-1.2ns_vs_nbv.inpcrd_ncca-fit @N,C,CA
trajout NBV_av_0.85-1.2ns_fit_nbv.inpcrd

The first couple of rms lines result in a trajout that's unchanged from the
trajin (so I didn't bother to copy it), and the next does an rms fitting to
the file nbv.inpcrd, similar I suppose to the VMD RMS Alignment program.

Here are the heads for the respective three molecule files (nbv.inpcrd is
the same as nbv.inpcrd.mdcrd, except that the former is a .restrt format
file and the latter has been changed by ptraj to a .mdcrd file for viewing
in VMD):

command: head -3 NBV_av_0.85-1.2ns nbv.inpcrd.mdcrd
NBV_av_0.85-1.2ns_fit_nbv.inpcrd
==> NBV_av_0.85-1.2ns <==
rdparm transformed trajectory
  21.787 -1.122 23.172 22.017 -1.201 22.212 22.614 -1.157 23.710
21.234
  -1.909 23.393 21.039 0.126 23.433 20.116 0.080 22.875 20.685
0.227

==> nbv.inpcrd.mdcrd <==
rdparm transformed trajectory
  21.614 -1.143 22.977 21.719 -1.208 21.954 22.544 -1.165 23.419
21.054
  -1.939 23.315 20.928 0.132 23.326 19.948 0.134 22.848 20.543
0.181

==> NBV_av_0.85-1.2ns_fit_nbv.inpcrd <==
rdparm transformed trajectory
   6.612 5.663 10.861 6.842 5.585 9.901 7.440 5.630 11.399
6.061
   4.874 11.082 5.861 6.909 11.123 4.938 6.861 10.565 5.507
7.009

        Obviously the first two are pretty similar and the third one is
transformed. But from the ptraj script, it would seem to me that they
should all be pretty similar in their coordinates; I don't know why ptraj
puts the coordinates for the third so different when it tries to do its rms
fitting. Doubtless that's why the total RMSD seen at the bottom of the VMD
calculator is so high (16.889...). Finally, I used the RMSD Alignment
program and moved the last molecule so that it pretty well overlaps the
other two, and then did a re-calculation in the RMSD Calculator. Not
surprisingly the rmsd was now low (0.9642...). But the ptraj rms value
from the above script was a bit higher (1.59172); I don't know why:

command: more rms_av_0.85-1.2ns_vs_nbv.inpcrd_all
    1.00 1.59172

(The value for rms_av_0.85-1.2ns_vs_nbv.inpcrd_all-fit was identical.)
Well, I realize that this question has diverged a bit into being a ptraj
question as well, but if anyone has any insight, I'd appreciate it.

Much thanks,
James Cavenaugh

----- End forwarded message ----- 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349              
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078