- Original Message -
From: Bernhard Knapp bernhard.kn...@meduniwien.ac.at
Date: Thursday, August 12, 2010 23:28
Subject: [gmx-users] Question regarding tpr files and rmsd
To: gmx-users@gromacs.org
Dear users
Due to a hard disk crash we lost several md simulations.
Fortunalty we have backup copies of the trajectry files
(xtc format) and structure files of the first frame of the
simulation (created via trjconv -b 0 -e 0 -f myName.md.trr
-o myName.md.firstframe.pdb -s myName.md.tpr). We do
not have the tpr files which we usually used for example for the
rmsd calculations. We found that the -s option of g_rms
does not only take tpr files but also pdb files, however
when I compare the resulting xvg files the values are
slightly different and I get a warning Warning: if there
are broken molecules in the trajectory file, they can not be
made whole without a run input file.
Sure. GROMACS-written PDB files contain no connectivity data. .tpr files do.
The average
difference (over 10 ns) between the xvg file based on the
tpr and on the pdb is 0.01447971.
Probably quite a reasonable difference in the average RMSD (in nm), given the
pdb approximates the floating-point numbers in the .tpr with only 3 decimal
points (in Angstrom)
Example for the 2 file:
[bkn...@quovadis02 test]$ sdiff rmsd.pdb.xvg rmsd.tpr.xvg | less
# This file was created Thu Aug 12 11:23:09
2010 | # This file was created Thu Aug 12 11:22:06 2010
# by the following
command: # by the following command:
# g_rms -f 1mi5_A6D.md.xtc -s 1mi5_A6D.md.firstframe.pdb -o r |
# g_rms -f 1mi5_A6D.md.xtc -s 1mi5_A6D.md.tpr -o rmsd.tpr.xvg
# #
# g_rms is part of G R O M A C
S: # g_rms is part of G R O M A C S:
# #
# GROtesk MACabre and
Sinister| # S C A M O R G
# #
@title
RMSD @title RMSD
@xaxis label Time
(ps) @xaxis label Time (ps)
@yaxis label RMSD
(nm) @yaxis label RMSD (nm)
@TYPE
xy@TYPE xy
@ subtitle Protein after lsq fit to
Protein @ subtitle Protein after lsq fit to Protein
0.000
0.0005041 |0.0000.0005046
3.000
0.1072081 |3.0000.0981387
6.000
0.1281023 |6.0000.1207779
9.000
0.1452615 |9.0000.1351306
...
My questions are now:
- Why are the xvg files different if they are based on the tpr
and on the pdb file?
PDB has more limited precision than the .tpr, and if you're fitting to their
structure, you'll get slightly different results.
- What is the more appropriate way to calculate the rmsd?
Depends what you're trying to measure - but I'd argue there shouldn't be a
significant difference between these two.
- Just if the more appropriate way is the tpr file: Is it valid
to recreate the tpr file via grompp solely on the firstframe.pdb
and the xtc of the trajectory? eg via
It's about the best you can do. IIRC, the firstframe.pdb will have higher
precision than the first frame of the .xtc, typically, and so produce a .tpr
closer to the original.
The procedure below is unnecessarily regenerating water. Use pdb2gmx to
generate the .top, and then simply use grompp to combine .mdp, .top and
whichever source of the first frame you choose.
pdb2gmx -f 1mi5_A6D.md.firstframe.pdb -o
1mi5_A6D.md.firstframe.pdb.gro -p 1mi5_A6D.md.top
editconf -f 1mi5_A6D.md.firstframe.pdb.gro -o
1mi5_A6D.firstframe.cube.pdb -bt cubic -d 2.0
genbox -cp 1mi5_A6D.firstframe.cube.pdb -cs spc216.gro -o
1mi5_A6D.firstframe.water.pdb -p 1mi5_A6D.md.top
grompp -f md.mdp -c 1mi5_A6D.firstframe.water.pdb -p
1mi5_A6D.md.top -o 1mi5_A6D.md.RECREATED.tpr
then the average difference between the xvg file based on the
tpr and the recreated tpr is 1.4865E-05 (which is much more
similar however still not identical)
You fitted and analyzed with respect to Protein, so the re-generation of the
water should have no effect here. The remaining difference should be
attributable to the difference between whatever coordinate changes pdb2gmx is
making (I can't tell what, if any) based on the two slightly different starting
points - but subjecting both to rounding to .gro precision. This last point
explains why they're so similar, I expect.
Mark
example:
[bkn...@quovadis02 test]$ sdiff rmsd.tprRECREATED.xvg
rmsd.tpr.xvg
