<div dir="ltr">Dear cp2k users,<div><br></div><div>Since a few people asked me to share the tpr2psf script, here is the link: https://github.com/pmamonov/tpr2psf</div><div><br></div><div>Regards,</div><div>Peter<br><br>On Saturday, August 28, 2010 3:51:54 PM UTC+4, Peter wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;">Dear All,
<br>Some time ago I wrote a (python) code to convert GROMACS binary TPR
<br>files into CP2K readable PSF (topology) and PRM (FF parameters) files.
<br>Unfortunately I didn't achieve exact match between energies calculated
<br>by GROMACS and CP2K/FIST. Respective energies calculated by GROMACS
<br>and CP2K/FIST match in case of bonds, angles and sometimes dihedrals.
<br>Nonbonded interactions (Coulomb and VdW, both long range and 1-4) are
<br>different for most of tested systems. Gradients calculated by both
<br>codes for some large system correlate with covariance coefficient 0.4.
<br>At some point I considered this result satisfactory and moved on. For
<br>now I don't have an intention to further improve this conversion tool,
<br>but probably others will find it to be a good starting point for
<br>creating such a conversion tool or probably try it as-is. So I put an
<br>archive with script, auxillary library (modified libgmx wrapper by
<br>Martin Hofling [<a href="http://oldwww.gromacs.org/pipermail/gmx-developers/2009-March/003183.html" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Foldwww.gromacs.org%2Fpipermail%2Fgmx-developers%2F2009-March%2F003183.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHmVM2yFP9PmTgAdUykq24dlWOLew';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Foldwww.gromacs.org%2Fpipermail%2Fgmx-developers%2F2009-March%2F003183.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHmVM2yFP9PmTgAdUykq24dlWOLew';return true;">http://oldwww.gromacs.org/<wbr>pipermail/gmx-developers/
<br>2009-March/003183.html</a>]) and some sample files here:
<br><a href="http://groups.google.com/group/cp2k/web/gmx2psf.tar.bz2" target="_blank" onmousedown="this.href='http://groups.google.com/group/cp2k/web/gmx2psf.tar.bz2';return true;" onclick="this.href='http://groups.google.com/group/cp2k/web/gmx2psf.tar.bz2';return true;">http://groups.google.com/<wbr>group/cp2k/web/gmx2psf.tar.bz2</a>
<br>
<br>Best regards,
<br>Peter Mamonov
<br>Lomonosov Moscow State University
<br>Biological faculty
<br>biophysics dept.
<br>
<br>P.S. Here is a help message from script, which provides some
<br>additional info:
<br>
<br>ABOUT:
<br>This script  reads GROMACS .tpr file and produces .psf and .prm files
<br>readable by  CP2K[ <a href="http://cp2k.berlios.de/" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fcp2k.berlios.de%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNG8a04cBV6g-qX2gL0GLMujwBbc_g';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fcp2k.berlios.de%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNG8a04cBV6g-qX2gL0GLMujwBbc_g';return true;">http://cp2k.berlios.de/</a>] and probably some other MM
<br>software.
<br>This script uses modified version of GROMACS library python wrapper,
<br>written by Martin Hofling  [<a href="http://oldwww.gromacs.org/pipermail/gmx-developers/2009-March/003183.html" target="_blank" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Foldwww.gromacs.org%2Fpipermail%2Fgmx-developers%2F2009-March%2F003183.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHmVM2yFP9PmTgAdUykq24dlWOLew';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Foldwww.gromacs.org%2Fpipermail%2Fgmx-developers%2F2009-March%2F003183.html\46sa\75D\46sntz\0751\46usg\75AFQjCNHmVM2yFP9PmTgAdUykq24dlWOLew';return true;">http://oldwww.gromacs.org/<wbr>pipermail/gmx-
<br>developers/2009-March/003183.<wbr>html</a>].
<br>
<br>NB: This script IS NOT capable of exact conversion between tpr and psf/
<br>prm topology/force field representation. Respective energies
<br>calculated by GROMACS and CP2K/FIST match in case of bonds, angles and
<br>sometimes dihedrals. Nonbonded interactions (Coulomb and VdW, both
<br>long range and 1-4) are different for most of tested systems.
<br>Gradients calculated by both codes for some large system correlate
<br>with covariance coefficient 0.4.
<br>
<br>REQUIREMENTS:
<br>* GROMACS dynamic libraries (libgmx*.so)
<br>* Modified version of GromPy wrapper (distributed  with this script)
<br>
<br>USAGE: tpr2psf.py system.tpr
<br>Will produce system.psf and system.prm
<br>
<br>
<br></blockquote></div></div>