Difference between revisions of "GMIN"

From Docswiki
Jump to navigation Jump to search
import>Csw34
import>Cjf41
 
(78 intermediate revisions by 2 users not shown)
Line 1: Line 1:
  +
[[Image:Gmin.jpg|thumb|GMIN in all its glory|300px|right]]
GMIN page - very much under construction!
 
  +
'''GMIN page - very much under construction!'''
   
 
= Introduction =
 
= Introduction =
Line 19: Line 20:
 
to the latest minimum in the chain or vary freely. The program knows many different empirical potentials, and it is straightforward to add new systems. From version 2.2 basin-sampling thermodynamics has been added, and from version 2.3 parallel tempering basin-sampling and basin-hopping have been implemented with MPI.
 
to the latest minimum in the chain or vary freely. The program knows many different empirical potentials, and it is straightforward to add new systems. From version 2.2 basin-sampling thermodynamics has been added, and from version 2.3 parallel tempering basin-sampling and basin-hopping have been implemented with MPI.
   
To start a calculation you need a file called <tt>data</tt> in the current directory, along with a file called <tt>coords</tt> containing the initial coordinates, which can be random. If the ''SEED'' keyword is present in <tt>data</tt> you also need a file called <tt>seed</tt> containing the seed coordinates. Most output is written to stdout, although the file <tt>lowest</tt> is always created at the end of the run, containing the energies and geometries of the lowest few configurations found in the given run. The geometries are saved in XMakemol xyz format.
+
To start a calculation you need a file called <tt>data</tt> in the current directory, along with a file called <tt>coords</tt> containing the initial coordinates, which can be random. If the [[SEED]] keyword is present in <tt>data</tt> you also need a file called <tt>seed</tt> containing the seed coordinates. Most output is written to stdout, although the file <tt>lowest</tt> is always created at the end of the run, containing the energies and geometries of the lowest few configurations found in the given run. The geometries are saved in XMakemol xyz format.
   
  +
= Source code =
  +
  +
You can download the latest version of the source code for all our programs under the GPL [http://www-wales.ch.cam.ac.uk/svn.tar.bz2 here]. This tarball does '''not''' contain the CHARMM and AMBER source. If you have a license and want to use them with GMIN, please contact Professor Wales.
   
 
= Keywords =
 
= Keywords =
  +
  +
'''I'm still deciding how to break the keywords up - it is currently done alphabetically - but it might make sense to do it by usage i.e. keywords that work only with AMBER or CHARMM in different sections''':
  +
  +
=== AMBER keywords (old implementation) ===
  +
=== AMBER9 keywords ===
  +
=== CHARMM keywords ===
  +
=== AMH keywords ===
  +
  +
'''Feedback needed!'''--[[User:Csw34|Csw34]] 09:51, 27 July 2010 (UTC)
   
 
== The data file ==
 
== The data file ==
Line 29: Line 42:
 
The following keywords are recognised, where n and x are integer and real data, respectively.
 
The following keywords are recognised, where n and x are integer and real data, respectively.
   
  +
=== #-F ===
 
* [[2D]]: enforce two-dimensional `flatland'.
 
* [[2D]]: enforce two-dimensional `flatland'.
   
 
* [[A9DIHE]] ''resid atom1 atom2 atom3 atom4 step'': specifies LeaP should be used to take a step in the dihedral for the atoms specified. It is limited to dihedrals within a residue. This keyword is not yet fully functional.
 
* [[A9DIHE]] ''resid atom1 atom2 atom3 atom4 step'': specifies LeaP should be used to take a step in the dihedral for the atoms specified. It is limited to dihedrals within a residue. This keyword is not yet fully functional.
   
* [[A9INTE]]: specifies that after each quench that does not lead to an inversion of chirality, isomerisation of a peptide bond or cold fusion - the interaction enthalpy between a specified residue and the rest of the system should be calculated using the external script `AMBGMINintE.sh', and read back into GMIN. This is intended for use with protein/ligand systems where you are searching for low energy docked structures. As the total energy does not fully correlate with the protein/ligand interaction enthalpy, it is often useful to retain not only the lowest [[SAVE]] total energy structures, but also the lowest [[SAVEINTE]] interaction enthalpy structures.
+
* [[A9INTE]]: specifies that after each quench that does not lead to an inversion of chirality, isomerisation of a peptide bond or cold fusion - the interaction enthalpy between a specified residue and the rest of the system should be calculated using the external script 'AMBGMINintE.sh', and read back into GMIN.
   
To use this keyword, the `AMBGMINintE.sh' script (contained in the SVN repository in the <tt>SCRIPTS</tt> directory) must be present in the GMIN working directory. You should ensure that you have edited it to match the residue numbering of your system. You also need a full AMBER9+ installation with access to the `sander' executable. When using this keyword, an interaction enthalpy dump file is produced very [[DUMPINT]] steps, and at the end of the run, structural output files are produced for the [[SAVEINTE]] lowest interaction enthalpy geometries. After each quench, the structure with the current lowest interaction energy is dumped in pdb and rst format prefixed with `bestint.' to allow monitoring.
+
This is intended for use with protein/ligand systems where you are searching for low energy docked structures. As the total energy does not fully correlate with the protein/ligand interaction enthalpy, it is often useful to retain not only the lowest [[SAVE]] total energy structures, but also the lowest [[SAVEINTE]] interaction enthalpy structures. To use this keyword, the `AMBGMINintE.sh' script (contained in the SVN repository in the <tt>SCRIPTS</tt> directory) must be present in the GMIN working directory. You should ensure that you have edited it to match the residue numbering of your system. You also need a full AMBER9+ installation with access to the `sander' executable. When using this keyword, an interaction enthalpy dump file is produced very [[DUMPINT]] steps, and at the end of the run, structural output files are produced for the [[SAVEINTE]] lowest interaction enthalpy geometries. After each quench, the structure with the current lowest interaction energy is dumped in pdb and rst format prefixed with `bestint.' to allow monitoring.
   
 
* [[ACCEPTRATIO]] ''accrat'': ''accrat'' is the required acceptance ratio for the MC exploration of the transformed surface. For fixed temperature runs (the default) the maximum step size is adjusted to try and meet the requested value of ''accrat'' for a fixed maximum step size the temperature is adjusted instead. The default value of ''accrat'' is a half.
 
* [[ACCEPTRATIO]] ''accrat'': ''accrat'' is the required acceptance ratio for the MC exploration of the transformed surface. For fixed temperature runs (the default) the maximum step size is adjusted to try and meet the requested value of ''accrat'' for a fixed maximum step size the temperature is adjusted instead. The default value of ''accrat'' is a half.
   
* [[ACKLAND]] ''id'': specifies an Ackland embedded atom metal potential coded by Dr Mihai-Cosmin Marinica. ''id'' specifies the particular metal: 1 is ?, 2 is ?, 3 is ?, 4 is ?, 5 is iron, 6 is a different iron, 7 is tunsten. Positive values for ''id'' specify periodic boundary conditions, where box lengths must be specified by the [[PERIODIC]] keyword. Negative values for ''id'' specify a cluster calculation. A [[CUTOFF]] value can also be used for clusters.
+
* [[ACKLAND]] ''id'': specifies an Ackland embedded atom metal potential coded by Dr Mihai-Cosmin Marinica. ''id'' specifies the particular metal: 1 is ?, 2 is ?, 3 is ?, 4 is ?, 5 is iron, 6 is a different iron, 7 is tungsten. Positive values for ''id'' specify periodic boundary conditions, where box lengths must be specified by the [[PERIODIC]] keyword. Negative values for ''id'' specify a cluster calculation. A [[CUTOFF]] value can also be used for clusters.
   
\item {\it ALGLUE\/}: specifies a glue potential for aluminium.
+
* [[ALGLUE]]: specifies a glue potential for aluminium.
   
\item {\it AMBER\/}: specifies the AMBER force field (old implementation). See also {\it DIELEC}.
+
* [[AMBER]]: specifies the AMBER force field ('''WARNING: old implementation'''). See also [[DIELEC]].
   
\item {\it AMBER9 inpcrd inpcrdformat\/}: specifies a calculation with the interfaced
+
* [[AMBER9]] ''inpcrd inpcrdformat'': specifies a calculation with the interfaced version of the Amber 9 program package.
version of the Amber 9 program package. From this package the Amber force fields
 
are being used, with small modifications ({\it e.g.} smooth cut-offs).
 
Starting coordinates do not need to be specified in the {\it odata} file, they
 
are read from {\it inpcrd} instead (default {\it coords.inpcrd}), in Amber inpcrd
 
file format specified by the second optional argument {\it inpcrdformat}.
 
If the second argument is missing, it is assumed that {\it inpcrd} contains
 
only three columns with the xyz coordinates of all atoms, in the same order
 
as in the topology file. To start a run with this interface,
 
several auxiliary files are required in the same directory: input coordinate file
 
{\it coords.inpcrd}, parameter topology file {\it coords.prmtop},
 
input file to Amber containing force field specifications {\it min.in}, and, if
 
desired, a coordinate file different from {\it coords.inpcrd} containing
 
starting coordinates.
 
To turn on smooth cutoffs for the Generalised Born force fields, the keyword
 
{\it ifswitch=1} has to be used in the {\it \&cntrl} namelist block of {\it min.in}.
 
When using the {\it AMBER9} keyword, any calculated second derivatives will be
 
numerical. If one wants analytical second derivatives, the {\it NAB} keyword
 
should be used instead, with the same syntax.
 
Additional keywords for the AMBER 9 runs are {\it DUMPSTRUCTURES}, {\it AMBERMDSTEPS},
 
{\it LIGMOVE (0.0-1.0) (x.x)} and {\it MOVABLEATOMS}.
 
% missing AMH - put it in after merger
 
   
  +
From this package the Amber force fields are being used, with small modifications (''e.g.'' smooth cut-offs). Starting coordinates are read from ''inpcrd'' instead (default ''coords.inpcrd''), in Amber inpcrd file format specified by the second optional argument ''inpcrdformat''. If the second argument is missing, it is assumed that ''inpcrd'' contains only three columns with the xyz coordinates of all atoms, in the same order as in the topology file. To start a run with this interface, several auxiliary files are required in the same directory: input coordinate file ''coords.inpcrd'', parameter topology file ''coords.prmtop'', input file to Amber containing force field specifications ''min.in'', and, if desired, a coordinate file different from ''coords.inpcrd'' containing starting coordinates. To turn on smooth cutoffs for the Generalised Born force fields, the keyword ''ifswitch=1'' has to be used in the ''\&cntrl'' namelist block of ''min.in''. When using the [[AMBER9]] keyword, any calculated second derivatives will be numerical. If one wants analytical second derivatives, the [[NAB]] keyword should be used instead, with the same syntax. Additional keywords for the AMBER 9 runs are [[DUMPSTRUCTURES]], [[AMBERMDSTEPS]], [[LIGMOVE]] ''(0.0-1.0) (x.x)'' and [[MOVABLEATOMS]].
\item {\it AMCHNMAX\/}: The maximum number of angles that will be changed by up to {\it STEP\/} during an
 
AMBER dihedral step. If this is not set or is set to zero, cartesian steps of maximum size {\it STEP\/} are taken
 
instead.
 
   
  +
* '''NO KEYWORD INFO FOR AMH :(''' - Mike?
\item {\it AMCHNMIN\/}: The minimum number of angles that will be changed during an AMBER dihedral step.
 
   
\item {\it AMCHPMAX\/}: The maximum probability for a single angle to be twisted in an AMBER dihedral step.
+
* [[AMCHNMAX]]: The maximum number of angles that will be changed by up to [[STEP]] during an AMBER dihedral step. If this is not set or is set to zero, cartesian steps of maximum size [[STEP]] are taken instead.
   
\item {\it AMCHPMIN\/}: The minimum probability for a single angle to be twisted in an AMBER dihedral step.
+
* [[AMCHNMIN]]: The minimum number of angles that will be changed during an AMBER dihedral step.
   
  +
* [[AMCHPMAX]]: The maximum probability for a single angle to be twisted in an AMBER dihedral step.
\item {\it ANGSTROM\/}: specifies coordinates in \AA ngstrom for the {\it FRAUSI\/}
 
potential.
 
   
  +
* [[AMCHPMIN]]: The minimum probability for a single angle to be twisted in an AMBER dihedral step.
\item {\it ARGON\/}: introduces a diatomics-in-molecules calculation for
 
a neutral, cationic or electronically excited argon cluster. See also
 
{\it GROUND\/}, {\it PLUS\/}, {\it TWOPLUS\/} and {\it STAR\/}.
 
   
  +
* [[ANGSTROM]]: specifies coordinates in <math>{\rm \AA}</math>ngstrom for the [[FRAUSI]] potential.
\item{\it ARM arma armb}: use the acceptance-ratio method (Bouzida et al., {\it Phys.~Rev.~A},
 
{\bf 45}, 8894, 1992) to adjust the step size to achieve the requested
 
acceptance ratio. A scaling factor is calculated and applied to {\it step}, {\it rotmax},
 
and/or {\it transmax}. The scaling factor is calculated according to
 
$\log(arma*P_{\rm t}+armb)/\log(arma*P_0+armb)$, where $P_{\rm t}$ defines the
 
target acceptance ratio and $P_0$ the actual acceptance ratio. Both values {\it arma} and
 
{\it armb} default to 0.4.
 
   
  +
* [[ARGON]]: introduces a diatomics-in-molecules calculation for a neutral, cationic or electronically excited argon cluster. See also [[GROUND]], [[PLUS]], [[TWOPLUS]] and [[STAR]].
\item {\it ARNO\/}: specifies a diatomics-in-molecules potential for Ar$_N$-NO clusters.
 
   
  +
* [[ARM]] ''arma armb'': use the acceptance-ratio method (Bouzida et al., ''Phys.Rev.A'', '''45''', 8894, 1992) to adjust the step size to achieve the requested acceptance ratio.
\item {\it AVOID dist maxsave}: specifies that the geometry should be reseeded if the
 
latest structure gets within a distance {\it dist} of the {\it maxsave} members of a
 
cyclic list.
 
   
  +
A scaling factor is calculated and applied to [[STEP]], ''rotmax'', and/or ''transmax''. The scaling factor is calculated according to <math>\log(arma*P_{\rm t}+armb)/\log(arma*P_0+armb)</math>
\item {\it AXTELL zstar\/}: specifies an additive Axilrod-Teller term for certain
 
  +
where <math>P_{\rm t}</math> defines the target acceptance ratio and <math>P_0</math> the actual acceptance ratio. Both values ''arma'' and
diatomics-in-molecules potentials as well as the Pacheco-Ramelho intermolecular potential for
 
  +
''armb'' default to 0.4.
C$_{60}$.\cite{pachecor97}
 
{\it zstar\/} is the coefficient multiplying this term.
 
   
  +
* [[ARNO]]: specifies a diatomics-in-molecules potential for Ar<math>_N</math>-NO clusters.
\item {\it BASIN bgmax\/}: specifies a basin-hopping run (as opposed to standard MC
 
on the untransformed surface). {\it bgmax\/} is the convergence threshold
 
on the RMS force in the basin-hopping
 
quenches. If this criterion is too strict then the run time will be greatly increased.
 
If it is too sloppy then the performance of the algorithm is impaired. Different values
 
are needed for different potentials. {\it SLOPPYCONV} can be used instead.
 
   
  +
* [[AVOID]] ''dist maxsave'': specifies that the geometry should be reseeded if the latest structure gets within a distance ''dist'' of the ''maxsave'' members of a cyclic list.
\item {\it BFGS}: specifies that the full BFGS minimiser should be used. Inefficient compared to LBFGS.
 
   
  +
* [[AXTELL]] ''zstar'': specifies an additive Axilrod-Teller term for certain diatomics-in-molecules potentials as well as the Pacheco-Ramelho intermolecular potential for C<math>_{60}</math>.<ref name="pachecor97">
\item {\it BHPT pttmin pttmax exchprob\/}: specifies minimum ({\it pttmin\/}) and maximum
 
  +
<bibtex>
({\it pttmax\/}) temperatures
 
  +
@Article{pachecor97,
for a parallel tempering basin-hopping run and the probability of attempting replica
 
  +
author = {Pacheco, J. M. and Ramalho, J. P. P.},
exchange ({\it exchprob\/}). Should be used together with the {\it MPI\/} keyword.
 
  +
title = {First-principles determination of the dispersion interaction between
(Only available if the source is compiled with MPI enabled.)
 
  +
fullerenes and their intermolecular potential},
  +
journal = {\prl},
  +
year = {1997},
  +
volume = {79},
  +
pages = {3873}
  +
}
  +
</bibtex>
  +
</ref>''zstar'' is the coefficient multiplying this term.
  +
  +
* [[BASIN]] ''bgmax'': specifies a basin-hopping run (as opposed to standard MC on the untransformed surface). ''bgmax'' is the convergence threshold on the RMS force in the basin-hopping quenches. If this criterion is too strict then the run time will be greatly increased. If it is too sloppy then the performance of the algorithm is impaired. Different values are needed for different potentials. [[SLOPPYCONV]] can be used instead.
  +
  +
* [[BFGS]]: specifies that the full BFGS minimiser should be used. Inefficient compared to LBFGS.
  +
  +
* [[BHPT]] ''pttmin pttmax exchprob'': specifies minimum (''pttmin'') and maximum (''pttmax'') temperatures for a parallel tempering basin-hopping run and the probability of attempting replica exchange (''exchprob''). Should be used together with the [[MPI]] keyword. (Only available if the source is compiled with MPI enabled.)
  +
  +
* [[BINARY]] ''ntypea epsab epsbb sigmaab sigmabb'': specifies a binary Lennard-Jones system. ''ntypea'' is the number of type A atoms - the rest are assumed to be type B and appear at the end of the list of coordinates. <math>\epsilon_{\rm AA}=\sigma_{\rm AA}=1</math> define the units of energy and length, and ''epsab''=<math>\epsilon_{\rm AB}</math>, ''epsbb''=<math>\epsilon_{\rm BB}</math>, ''sigmaab''=<math>\sigma_{\rm AB}</math>, ''sigmabb''=<math>\sigma_{\rm BB}</math>. The box parameters and cutoff should be specified with the [[PERIODIC]] keyword.
   
  +
* [[BINSTRUCTURES]] ''SaveNth'': requests that the geometry of every ''SaveNth'' new structure found during basin-sampling is recorded in <tt>binstructures.j</tt>, where ''j'' is the index of the bin to which a given minimum belongs. If this keyword is present then GMIN switches from plain [[PTMC]] to [[BSPT]]. Without [[BINSTRUCTURES]] the [[BSPT]] keyword will perform a standard PTMC run with no quenching.
\item {\it BINARY ntypea epsab epsbb sigmaab sigmabb\/}: specifies a binary Lennard-Jones
 
system. {\it ntypea\/} is the number of type
 
A atoms---the rest are assumed to be type B and appear at the end of the list
 
of coordinates. $\epsilon_{\rm AA}=\sigma_{\rm AA}=1$ define the units of energy and length,
 
and {\it epsab\/}=$\epsilon_{\rm AB}$, {\it epsbb\/}=$\epsilon_{\rm BB}$,
 
{\it sigmaab\/}=$\sigma_{\rm AB}$, {\it sigmabb\/}=$\sigma_{\rm BB}$.
 
The box parameters and cutoff should be specified with the {\it PERIODIC\/} keyword.
 
   
  +
* [[BLJCLUSTER]] ''ntypea epsab epsbb sigmaab sigmabb cutoff'': specifies a binary Lennard-Jones cluster. The parameters are the same as for [[BINARY]], above.
\item {\it BINSTRUCTURES SaveNth}: requests that the geometry of every {\it SaveNth}
 
new structure found during basin-sampling is
 
recorded in {\tt binstructures.j}, where {\it j} is the index of the bin
 
to which a given minimum belongs. If this keyword is
 
present then GMIN switches from plain PTMC to BSPT.
 
Without {\it BINSTRUCTURES} the {\it BSPT} keyword will perform a
 
standard PTMC run with no quenching.
 
   
  +
* [[BLN]] <math>k_r</math> <math>k_\theta</math>: specifies a BLN off-lattice protein model with bond-length and bond-angle force constants <math>k_r</math> and <math>k_\theta</math>. An auxiliary file ''BLNsequence'' is required.
\item {\it BLJCLUSTER ntypea epsab epsbb sigmaab sigmabb cutoff\/}: specifies a binary Lennard-Jones
 
cluster. The parameters are the same as for {\it BINARY\/}, above.
 
   
  +
* [[BSMIN]]: specifies a Bulirsch-Stoer minimisation scheme. Very inefficient compared to LBFGS.
\item {\it BLN $k_r$ $k_\theta$ \/}: specifies a BLN off-lattice protein model with
 
bond-length and bond-angle force constants $k_r$ and $k_\theta$.
 
An auxiliary file {\tt BLNsequence} is required.
 
See \S \ref{sec:BLN} for more details.
 
   
  +
* [[BSPT]] ''histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nquench nenrper hbins qfrq'': requests a basin-sampling run to accumulate the quench probability for local minima as a function of potential energy using a parallel-tempering algorithm.
\item {\it BSMIN\/}: specifies a Bulirsch-Stoer minimisation scheme.
 
Very inefficient compared to LBFGS.
 
   
  +
This keyword also specifies the energy range for the histogram of quench energies, ''histmin'' to ''histmax'', the energy range for the histogram of instantaneous configurations, ''ptemin'' to ''ptemax'', the temperature range (''pttmin'' and ''pttmax''), the probability of attempting an exchange ''exchprob'', the number of equilibration steps, ''nequil'', the number of parallel tempering MC steps without quenching, ''ptsteps'', the number of parallel tempering MC steps with quenching, ''nquench'', the number of bins for the histogram of instantaneous potential energy, ''nenrper'', the number of bins for the histogram of quench energies, ''hbins'', and the quench frequency, ''qfrq''. Should be used together with the [[MPI]] keyword. (This option is only available if the source is compiled with an MPI enabled.)
\item {\it BSPT histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nquench nenrper hbins qfrq\/}:
 
requests a basin-sampling run to accumulate the quench probability for local minima
 
as a function of potential energy using
 
a parallel-tempering algorithm.
 
This keyword also specifies the energy range for the histogram of quench energies,
 
{\it histmin\/} to {\it histmax\/},
 
the energy range for the histogram of instantaneous configurations, {\it ptemin} to {\it ptemax},
 
the temperature range ({\it pttmin} and {\it pttmax}),
 
the probability of attempting an exchange {\it exchprob}, the
 
number of equilibration steps, {\it nequil},
 
the number of parallel tempering MC steps without quenching, {\it ptsteps},
 
the number of parallel tempering MC steps with quenching, {\it nquench},
 
the number of bins for the histogram of instantaneous potential energy, {\it nenrper},
 
the number of bins for the histogram of quench energies, {\it hbins},
 
and the quench frequency, {\it qfrq}.
 
Should be used together with the {\it MPI\/} keyword. % and {\it BINSTRUCTURES\/} keywords.
 
(This option is only available if the source is compiled with an MPI enabled.)
 
   
\item {\it BSPTDUMPFRQ n\/}, {\it n\/} is the interval at which intermediate statistics
+
* [[BSPTDUMPFRQ]] ''n'': ''n'' is the interval at which intermediate statistics and [[BSPTRESTART]] files are dumped. If ''n'' is less than one these files will only be dumped at the end of a complete run. See also [[BSPTRESTART]]. '''Note BSPTRESTART is NOT documented'''.
and {\it bsptrestart\/} files are dumped. If {\it n\/} is less than one these files
 
will only be dumped at the end of a complete run.
 
See also {\it BSPTRESTART\/}.
 
   
  +
* [[BSPTDUMPFRQ]]: restart a previous [[BSPT]] or [[PTMC]] run. The instantaneous and quench potential energy histograms are read from the last <tt>Visits.his</tt> and <tt>Visits2.his</tt> files, and the current state from <tt>bsptrestart</tt> files (one per node, numbered from zero). A finished run can be continued with more steps by changing the ''nquench'' or ''ptsteps'' parameters on the [[BSPT]] or [[PTMC]] line of the data file. Setting the interval for [[BSPTDUMPFRQ]] to
\item {\it BSPTDUMPFRQ\/}: restart a previous {\it BSPT\/} or {\it PTMC\/} run.
 
The instantaneous and quench potential energy histograms are read from the last
 
{\tt Visits.his} and {\tt Visits2.his} files, and the current state from
 
{\tt bsptrestart} files (one per node, numbered from zero).
 
A finished run can be continued with more steps by changing the {\it nquench}
 
or {\it ptsteps} parameters on the {\it BSPT\/} or {\it PTMC\/} line of
 
the data file. Setting the interval for {\it BSPTDUMPFRQ} to
 
 
minus one will read the last set of dump files.
 
minus one will read the last set of dump files.
   
% \item {\it BSWL\/}: obsolete Wang-Landau basin-sampling; do not use.
+
* [[BSWL]]: obsolete Wang-Landau basin-sampling; '''do not use'''.
   
\item {\it CAPSID rho epsilon radius height\/}: specifies a coarse-grained potential to represent virus capsid pentamers
+
* [[CAPSID]] ''rho epsilon radius height'': specifies a coarse-grained potential to represent virus capsid pentamers with parameters <math>\rho</math>, <math>\epsilon_2</math>, <math>r</math> and <math>height</math>, respectively. If <math>height</math> is omitted the default is 0.5.
with parameters $\rho$, $\epsilon_2$, $r$ and $height$, respectively.
 
If $height$ is omitted the default is 0.5.
 
   
\item {\it CENTRE \/}: if present the system will be translated so that the centre-of-mass
+
* [[CENTRE]]: if present the system will be translated so that the centre-of-mass lies at the origin after every quench.
lies at the origin after every quench.
 
   
\item {\it CENTREXY \/}: if present the system will be translated so that the centre-of-mass
+
* [[CENTREXY]]: if present the system will be translated so that the centre-of-mass lies at the centre of the xy plane after every quench. This is useful when using an implicit membrane like IMM1 where you have directionality only in the z-direction, so centreing in x and y should have no delaterious effect.
lies at the centre of the xy plane after every quench. This is useful when using an implicit membrane like IMM1 where you have directionality only in the
 
z-direction, so centreing in x and y should have no delaterious effect.
 
   
\item {\it CG \/}: specifies a conjugate-gradient minimisation scheme. Inefficient compared to LBFGS.
+
* [[CG]]: specifies a conjugate-gradient minimisation scheme. Inefficient compared to LBFGS.
   
\item {\it CHANGEACCEPT naccept\/}: {\it naccept\/} is an integer which sets the interval
+
* [[CHANGEACCEPT]] ''naccept'': ''naccept'' is an integer which sets the interval at which the acceptance ratio is checked and possible adjustments are made to the maximum step size or the temperature. The default is ''naccept''<math>=50</math>.
at which the acceptance ratio is checked and possible adjustments are made to the maximum
 
step size or the temperature. The default is {\it naccept\/}$=50$.
 
 
 
  +
* [[CHARMM]]: specifies that a CHARMM potential should be used. See also keywords [[CHARMMTYPE]], [[CHPMAX]], [[CHPMIN]], [[CHNMAX]], [[CHNMIN]], [[NOPHIPSI]], [[TOMEGA]], [[INTMIN]], [[CHFREQ]], [[CHRIGIDROT]], [[CHRIGIDTRANS]], and [[RMS]]. If [[CHNMAX]] is not specified, a cartesian displacement step taking scheme will be used. For cartesian steps, rings are moved as rigid bodies to avoid false knotted minima. See [[RINGROTSCALE]]. Finally, Molecular Dynamics (MD) can be employed to generate new geometries. See [[CHMD]] below.
\item {\it CHARMM}: specifies that a CHARMM potential should be used.
 
See also keywords {\it CHARMMTYPE}, {\it CHPMAX}, {\it CHPMIN}, {\it CHNMAX}, {\it CHNMIN},
 
{\it NOPHIPSI}, {\it TOMEGA}, {\it INTMIN}, {\it CHFREQ}, {\it CHRIGIDROT},
 
{\it CHRIGIDTRANS}, and {\it RMS}. If {\it CHNMAX} is not specified, a cartesian
 
displacement step taking scheme will be used. For cartesian steps, rings are moved as rigid bodies to avoid false knotted minima. See {\it RINGROTSCALE}. Finally, Molecular Dynamics (MD) can be employed to generate new geometries. See {\it CHMD}
 
   
\item {\it CHMD CHMDFREQ\/}: Requests Molecular Dynamics (MD) runs to be performed every {\it CHMDFREQ} step to generate new geometries. A {\it CHMDFREQ} setting of 20 will execute an MD run every 20$^\mathrm{th}$ step, while dihedral or cartesian moves are applied otherwise as specified in the data file. A CHARMM parameter file named 'chmd.par' containing all relevant keywords for the CHARMM {\it DYNA} module has to be present in the working directory. All CHARMM keywords must be uppercase and given in the first line. A typical example is:
+
* [[CHMD]] ''chmdfreq'': Requests Molecular Dynamics (MD) runs to be performed every ''chmdfreq'' steps to generate new geometries. A ''chmdfreq'' setting of 20 will execute an MD run every 20<math>^\mathrm{th}</math> step, while dihedral or cartesian moves are applied otherwise as specified in the data file. A CHARMM parameter file named <tt>chmd.par</tt> containing all relevant keywords for the CHARMM ''DYNA'' module has to be present in the working directory. All CHARMM keywords must be uppercase and given in the first line. A typical example is:
   
VERL NSTEP 500 TIMESTEP 0.002 TWINDH 10.0 IEQFRQ 200 ICHECW 1 IASORS 0 IASVEL 1 FIRS 500 FINA 500
+
''VERL NSTEP 500 TIMESTEP 0.002 TWINDH 10.0 IEQFRQ 200 ICHECW 1 IASORS 0 IASVEL 1 FIRS 500 FINA 500''
   
Please consult the CHARMM manual for further details on the {\it DYNA} module. Currently, the length of the input string given in 'chmd.par' is limited to 500 characters.
+
Please consult the CHARMM manual for further details on the ''DYNA'' module. Currently, the length of the input string given in <tt>chmd.par</tt> is limited to 500 characters.
   
\item{\it CHARMMENERGIES}: prints the components of the total CHARMM energy after each step.
+
* [[CHARMMENERGIES]]: prints the components of the total CHARMM energy after each step.
   
\item {\it CHARMMTYPE topfile paramfile\/}: {\it topfile} and {\it paramfile} are the
+
* [[CHARMMTYPE]] ''topfile paramfile'': ''topfile'' and ''paramfile'' are the common CHARMM top and param files, e.g. 'toph19\_eef1\_perm.inp' and 'param19\_eef1\_perm.inp'.
common CHARMM top and param files , e.g., `toph19\_eef1\_perm.inp' and `param19\_eef1\_perm.inp'.
 
   
\item{\it CHFREQ nfreq}: used with {\it CHARMM} keyword to specify that every
+
* [[CHFREQ]] ''nfreq'': used with [[CHARMM]] keyword to specify that every ''nfreq'' basin-hopping steps dihedrals are twisted. Default is ''nfreq''=1.
{\it nfreq} basin-hopping steps dihedrals are twisted. Default is {\it nfreq}=1.
 
   
\item{\it CHNMAX}: used with {\it CHARMM} keyword to specify the maximum allowed
+
* [[CHNMAX]]: used with [[CHARMM]] keyword to specify the maximum allowed number of angles to be twisted. Specifies a dihedral angle step taking scheme.
number of angles to be twisted. Specifies a dihedral angle step taking scheme.
 
   
\item{\it CHNMIN}: used with {\it CHARMM} keyword to specify the minimum allowed
+
* [[CHNMIN]]: used with [[CHARMM]] keyword to specify the minimum allowed number of angles to be twisted.
number of angles to be twisted.
 
   
\item{\it CHPMAX}: used with {\it CHARMM} keyword to specify the maximum allowed
+
* [[CHPMAX]]: used with [[CHARMM]] keyword to specify the maximum allowed probability for twisting an angle.
probability for twisting an angle.
 
   
\item{\it CHPMIN}: used with {\it CHARMM} keyword to specify the minimum allowed
+
* [[CHPMIN]]: used with [[CHARMM]] keyword to specify the minimum allowed probability for twisting an angle.
probability for twisting an angle.
 
   
  +
* [[CHRIGIDROT]] ''prot rotmax nrot'': used with [[CHARMM]] keyword to support rigid body rotation every ''nrot'' basin-hopping steps with maximum allowed probability ''prot'' and maximum allowed rotation angle ''rotmax'' (in degrees). The keyword [[CHRIGIDROT]] requires a file <tt>segments.tomove</tt>, which specifies the segments for rigid rotation. The segments are numbered and each line contains only one number.
\item{\it CHRIGIDROT prot rotmax nrot}: used with {\it CHARMM} keyword
 
to support rigid body rotation every {\it nrot} basin-hopping steps with maximum allowed
 
probability {\it prot} and maximum allowed rotation angle {\it rotmax} (in degrees).
 
The keyword {\it CHRIGIDROT} requires a file {\tt segments.tomove}, which specifies
 
the segments for rigid rotation. The segments are numbered and each line contains only one number.
 
   
  +
* [[CHRIGIDTRANS]] ''ptrans transmax ntrans'': used with [[CHARMM]] keyword to support rigid body translation every ''ntrans'' basin-hopping steps with maximum allowed probability ''ptrans'' and maximum allowed translation ''transmax'' (in <math>\AA</math>). The keyword [[CHRIGIDTRANS]] requires a file <tt>segments.tomove</tt>, which specifies the segments for rigid translation. The segments are numbered and each line contains only one number. [[CHRIGIDROT]] and [[CHRIGIDTRANS]] use the same <tt>segments.tomove</tt>.
\item{\it CHRIGIDTRANS ptrans transmax ntrans}: used with {\it CHARMM} keyword
 
to support rigid body translation every {\it ntrans} basin-hopping steps with maximum allowed
 
probability {\it ptrans} and maximum allowed translation {\it transmax} (in \AA).
 
The keyword {\it CHRIGIDTRANS} requires a file {\tt segments.tomove}, which specifies
 
the segments for rigid translation. The segments are numbered and each line contains only one number.
 
{\it CHRIGIDROT} and {\it CHRIGIDTRANS} use the same {\tt segments.tomove}.
 
   
  +
* [[CISTRANS]]: disables all checks for cis or deformed amide/peptide bonds.
% \item{\it NOCISTRANS}: not used
 
% \item{\it NORANDOM}: not used
 
% \item{\it PERMDIHE n1 n2 n3 etc.}: not used
 
   
  +
* [[COLDFUSION]] ''thresh'': if the energy falls below threshold ''thresh'' then cold fusion is assumed to have occurred and geometry optimisation stops. The default value is <math>-10^6</math>.
\item {\it CISTRANS\/}: disables all checks for cis or deformed amide/peptide bonds.
 
   
  +
* [[COMPRESS]] ''comp'': add a harmonic compression potential with force constant ''comp'' using the centre-of-mass distance for each atom.
\item {\it COLDFUSION thresh\/}: if the energy falls below threshold {\it thresh} then
 
cold fusion is assumed to have occurred and geometry optimisation stops.
 
The default value is $-10^6$.
 
   
  +
* [[COMMENT]]: the rest of the line is ignored.
\item {\it COMPRESS comp\/}: add a harmonic compression potential with force constant {\it comp\/} using the
 
centre-of-mass distance for each atom.
 
   
  +
* [[COOPMOVE]] ''n cut'': specifies cooperative moves in the step-taking routine. An atom is selected at random, and the ''n'' nearest neighbours (default 5) that lie within a cutoff distance of ''cut'' (default 1.0) are moved by the same amount.
\item {\it COMMENT \/}: the rest of the line is ignored.
 
   
  +
* [[CPMD]] ''sys'': specifies that the CPMD program should be called for energies and gradients. '''Not tested!'''
\item {\it COOPMOVE n cut\/}: specifies cooperative moves in the step-taking routine. An atom is
 
selected at random, and the {\it n} nearest neighbours (default 5) that lie within a cutoff
 
distance of {\it cut} (default 1.0) are moved by the same amount.
 
   
  +
* [[CUTOFF]] ''cutoff'': sets a cutoff beyond which the potential is truncated. This only has an effect for tight-binding silicon at present.
\item {\it CPMD sys\/}: specifies that the CPMD program should be called for energies and gradients. Not
 
tested!
 
   
  +
* [[DBRENT]] specifies minimisation using Brent's method with first derivatives in the conjugate-gradient procedure. Inefficient compared to LBFGS.
\item {\it CUTOFF cutoff\/}: sets a cutoff beyond which the potential is truncated. This
 
only has an effect for tight-binding silicon at present.
 
   
  +
* [[DEBUG]]: sets various debug printing options including the dumping of initial geometries and energies (to ''dump.X.xyz'') if [[DUMP]] is also set.
\item {\it DBRENT}: specifies minimisation using Brent's method with first derivatives in the
 
conjugate-gradient procedure.
 
Inefficient compared to LBFGS.
 
   
  +
* [[DECAY]] ''x'': magnitude of random move decays according to parameter ''x'' with distance from a randomly chosen atom.
\item {\it DEBUG\/}: sets various debug printing options including the dumping of initial
 
geometries and energies (to {\it dump.X.xyz\/}) if {\it DUMP} is also set.
 
   
  +
* [[DF1]]: specifies a binary 2D potential. The first <math>N/2</math> atoms have unit radius and the rest have radius 1.4, with a cutoff for each pair type at the average radius. The keyword [[2D]] must also be specified, along with a [[PERIODIC]] line to specify two box-lengths. Initial work uses box lengths of 3.31437171 for a number density of 0.9.
\item {\it DECAY x\/}: magnitude of random move decays according to parameter
 
{\it x\/} with distance from a randomly chosen atom.
 
   
\item {\it DF1\/}: specifies a binary 2D potential.
+
* [[DFTB]]: specifies a DFT-based tight-binding potential; the multiplicity is specified by keyword [[MULTIPLICITY]].
The first $N/2$ atoms have unit radius and the rest
 
have radius 1.4, with a cutoff for each pair type at the
 
average radius.
 
The keyword {\it 2D\/} must also be specified, along with a
 
{\it PERIODIC\/} line to specify two box-lengths.
 
Initial work uses box lengths of 3.31437171 for a number density of 0.9.
 
   
  +
* [[DGUESS]] ''dguess'': initial guess for diagonal elements of the inverse Hessian, used whenever the LBFGS optimiser is reset. The default is dguess=0.1.
\item {\it DFTB\/}: specifies a DFT-based tight-binding potential; the multiplicity is specified by
 
keyword {\it MULTIPLICITY\/}.
 
   
  +
* [[DIELEC]] ''dparam'': specifies dielectric constant for [[AMBER]] - '''old implementation'''.
\item {\it DGUESS dguess\/}: initial guess for diagonal elements of the inverse
 
Hessian, used whenever the LBFGS optimiser is reset.
 
The default is dguess=0.1.
 
   
  +
* [[DIPOLES]]: causes the first order induction energy to be included in a diatomics-in-molecules calculation for Ne<math>^+_n</math> or Ar<math>^+_n</math>. By default this term is neglected, although it may be significant.
\item {\it DIELEC dparam\/}: specifies dielectric constant for {\it AMBER\/}.
 
   
  +
* [[DUMP]]: if present will cause the energy and quench geometry for every step to be dumped into <tt>dump.X.xyz</tt> where ''X'' is an integer. The geometries are saved in XMakemol xyz format. If [[CHARMM]] is also specified, <tt>dump.pdb</tt> and <tt>dump.crd</tt> are produced containing each quench geometry in PDB and CHARMM CRD format.
\item {\it DIPOLES\/}: causes the first order induction energy to be included
 
in a diatomics-in-molecules calculation for Ne$^+_n$ or Ar$^+_n$. By default this
 
term is neglected, although it may be significant.
 
   
  +
* [[DUMPINT]] ''int'': changes the default interval for dumping a restart <tt>GMIN.dump</tt> file from 1000 basin-hopping steps to ''int''.
\item {\it DUMP\/}: if present will cause the energy and quench geometry for every step
 
to be dumped into {\it dump.X.xyz\/} where X is an integer. The geometries are saved
 
in XMakemol xyz format. If {\it CHARMM\/} is also specified, {\it dump.pdb\/} and {\it dump.crd\/}
 
are produced containing each quench geometry in PDB and CHARMM CRD format.
 
   
  +
* [[DUMPQU]]: when using [[AMBER9]], dumps each quench geometry in rst format to <tt>quenchX.rst</tt> and pdb format to <tt>quenchX.pdb</tt>. Dumping does not occur if a chirality check fails.
\item {\it DUMPINT int\/}: changes the default interval for dumping a restart
 
{\tt GMIN.dump} file from 1000 basin-hopping steps to {\it int\/}.
 
   
  +
* [[DZUGUTOV]] ''dzp1 dzp2 dzp3 dzp4 dzp5 dzp6 dzp7'': Dzugutov potential in a general form. The parameters are <math>m</math>, <math>A</math>, <math>aa</math>, <math>B</math>, <math>d</math>, <math>bb</math> and <math>m2</math>.
\item {\it DUMPQU\/}: when using {\it AMBER9\/}, dumps each quench geometry in rst format to quenchX.rst
 
and pdb format to quenchX.pdb. Dumping does not occur if a chirality check fails.
 
   
  +
* [[EAMAL]]: specifies an embedded atom model for aluminium.
\item {\it DZUGUTOV dzp1 dzp2 dzp3 dzp4 dzp5 dzp6 dzp7\/}: Dzugutov potential in a general form.
 
The parameters are $m$, $A$, $aa$, $B$, $d$, $bb$ and $m2$.
 
   
  +
* [[EAMLJ]] ''A0 beta Z0'': specifies the EAMLJ potential (Baskes, ''Phys.Rev.Lett.'', '''27''', 2592, 1999) with parameters ''A0'', ''beta'' and ''Z0''.
\item {\it EAMAL}: specifies an embedded atom model for aluminium.
 
   
  +
* [[EDIFF]] ''econv'': quench minima are only considered to be different if their energies differ by at least <math>econv</math>. This option mainly affects the lowest energy saved geometries. If the current quench energy is within <math>econv</math> of a saved energy, but lies lower, then the saved energy and geometry are replaced. The default is <math>0.02</math> but different values are appropriate for different potentials.
\item {\it EAMLJ A0 beta Z0\/}: specifies the EAMLJ potential (Baskes, {\it Phys.~Rev.~Lett.\/},
 
{\bf 27}, 2592, 1999) with parameters {\it A0\/}, {\it beta\/} and {\it Z0\/}.
 
   
  +
* [[EQUILIBRATION]] ''equil DumpEveryNthQuench'': ''equil'' is the number of MC steps preceding the accumulation of the density of states histogram in a Wang-Landau basin-sampling run. The default is 0. ''DumpEveryNthQuench'' specifies how often the statistics are recorded into the output files.
\item {\it EDIFF econv\/}: quench minima are only considered to be different if their
 
energies differ by at least $econv$. This option mainly affects the lowest energy
 
saved geometries. If the current quench energy is within $econv$ of a saved energy, but
 
lies lower, then the saved energy and geometry are replaced.
 
The default is $0.02$ but different values are appropriate for different potentials.
 
   
  +
* [[FAKEWATER]]: specifies a distance-dependent dielectric in [[AMBER]] - '''old implementation'''.
\item {\it EQUILIBRATION equil DumpEveryNthQuench\/}: {\it equil} is the number of
 
MC steps preceding the accumulation of the
 
density of states histogram in a Wang-Landau
 
basin-sampling run. The default is 0. {\it DumpEveryNthQuench} specifies how often the
 
statistics are recorded into the output files.
 
   
  +
* [[FIXEDEND]]: '''requires documentation :('''.
\item {\it FAKEWATER \/}: specifies a distance-dependent dielectric in {\it AMBER\/}.
 
   
  +
* [[FAL]]: specifies the Farkas potential for aluminium.
\item {\it FIXEDEND \/}: requires documentation.
 
   
  +
* [[FIXBOTH]]: both the temperature and maximum step size are fixed regardless of the calculated acceptance ratio.
\item {\it FAL \/}: specifies the Farkas potential for aluminium.
 
   
  +
* [[FIXCOM]]: fix centre of mass rather than centre of coordinates.
\item {\it FIXBOTH \/}: both the temperature and maximum step size are fixed regardless of
 
the calculated acceptance ratio.
 
   
  +
* [[FIXSTEP]]: the maximum step size is fixed and the temperature is varied to try and achieve the requested acceptance ratio.
\item {\it FIXCOM \/}: fix centre of mass rather than centre of coordinates.
 
   
  +
* [[FIXTEMP]]: explicitly fixes the temperature. Only used if [[FIXSTEP]] is set, in which case using [[FIXTEMP]] gives a result equivalent to [[FIXBOTH]].
\item {\it FIXSTEP \/}: the maximum step size is fixed and the temperature is varied to
 
try and achieve the requested acceptance ratio.
 
   
  +
* [[FNI]]: specifies the Farkas potential for nickel.
\item {\it FIXTEMP \/}: explicitly fixes the temperature. Only used if {\it FIXSTEP\/} is set, in
 
which case using {\it FIXTEMP\/} gives a result equivalent to {\it FIXBOTH\/}.
 
   
\item {\it FNI \/}: specifies the Farkas potential for nickel.
+
* [[FRAUSI]]: specifies a particular tight-binding potential for silicon. See also keyword [[ANGSTROM]].
   
  +
* [[FREEZE]] ''n1 n2'' <math>\ldots</math>: freeze the coordinates of atoms ''n1, n2,''<math>\ldots</math>}. Note that all the ''FREEZE'' keywords do '''NOT''' function when using [[AMBERMDSTEPS]]. Use multiple FREEZE statements for larger numbers of atoms.
\item {\it FRAUSI \/}: specifies a particular tight-binding potential for silicon.
 
See also keyword {\it ANGSTROM\/}.
 
   
\item {\it FREEZE n1 n2 $\ldots$ \/}: freeze the coordinates of atoms {\it n1, n2,$\ldots$}. Note that all the {\it FREEZE\/} keywords do NOT function
+
* [[FREEZERES]] ''n1 n2'' <math>\ldots</math>: freeze the coordinates of all atoms in residues ''n1, n2,''<math>\ldots</math>.
when using AMBERMDSTEPS.
 
   
\item {\it FREEZERES n1 n2 $\ldots$ \/}: freeze the coordinates of all atoms in residues {\it n1, n2,$\ldots$}.
+
* [[FREEZEALL]] ''n1 n2'' <math>\ldots</math>: freeze the coordinates of all atoms.
   
  +
* [[FS]] ''gatom'': specifies a Finnis-Sinclair potential using parameters from Finnis and Sinclair, ''Phil.Mag.A'', '''50''', 45 (1984) and corresponding erratum ''Phil.Mag.A'', '''53''', 161 (1986). ''gatom''=1 for V, ''gatom''=2 for Nb, ''gatom''=3 for Ta, ''gatom''=4 for Cr, ''gatom''=5 for Mo, ''gatom''=6 for W, ''gatom''=7 for Fe (original parameters), ''gatom''=8 for Fe (modified parameters in erratum). Subtoutine '''FS''' was coded by James Elliott in April 2009.
\item {\it FREEZEALL n1 n2 $\ldots$ \/}: freeze the coordinates of all atoms
 
   
  +
=== G-L ===
\item {\it UNFREEZE n1 n2 $\ldots$ \/}: unfreeze the coordinates of atoms {\it n1, n2,$\ldots$}. Only functions in conjunction with
 
  +
* [[GROUND]]: when combined with keywords [[NEON]] or [[ARGON]] uses an accurate (Aziz) potential to model the ground state neutral cluster.
{\it FREEZEALL\/}
 
   
  +
* [[GUIDE]] ''guidecut'': specifies the RMS force below which the real potential is used rather than a guiding potential. The systems affected are [[CPMD]] and [[WELCH]], which are guided by [[AMBER]] and [[TOSI]], respectively, and also [[PACHECO]], where the Axilrod-Teller contribution is only included when the RMS force falls below ''guidecut''. Default ''guidecut''=0.0001. New guided potentials are [[ZETT1]] and [[ZETT2]] (guided by Morse with <math>\rho=5</math>) and [[NATB]] (guided by [[GUPTAT]]). Parameters for the guiding potential must also be specified in <tt>data</tt>.
\item {\it UNFREEZERES n1 n2 $\ldots$ \/}: unfreeze the coordinates of all atoms in residues {\it n1, n2,$\ldots$}. Only functions in conjunction with
 
{\it FREEZEALL\/}
 
   
  +
* [[GUPTA]] ''gatom'': specifies a Gupta potential using parameters from Cleri and Rosato, ''Phys.Rev.B'', '''48''', 22 (1993). ''gatom''=1 for Ni, ''gatom''=2} for Cu, ''gatom''=3 for Rh, ''gatom''=4 for Pd, ''gatom''=5 for Ag, ''gatom''=6 for Ir, ''gatom''=7 for Pt, ''gatom''=8 for Au, ''gatom''=9 for Al, ''gatom''=10 for Pb, ''gatom''=11 for Ti type 1, ''gatom''=12 for Ti type 2, ''gatom''=13 for Zr type 1, ''gatom''=14 for Zr type 2, ''gatom''=15} for Co, ''gatom''=16 for Cd type 1, ''gatom''=17 for Cd type 2, ''gatom''=18 for Zn, ''gatom''=19 for Mg, ''gatom''=20 for V, ''gatom''=21 for Na, ''gatom''=22 for Sr (Wang and Blaisten-Barojas, ''J.Chem.Phys.'', '''115''', 3640 (2001)), ''gatom''=? for Au as used by Garzon et al. The '''Gupta''' subroutine was recoded more efficiently by James Elliott in April 2009.
\item{\it FS gatom}: specifies a Finnis-Sinclair potential using parameters from
 
Finnis and Sinclair, {\it Phil.~Mag.~A}, {\bf 50}, 45 (1984)
 
and corresponding erratum {\it Phil.~Mag.~A}, 53, 161 (1986).
 
{\it gatom}=1 for V, {\it gatom}=2 for Nb, {\it gatom}=3 for Ta, {\it gatom}=4
 
for Cr, {\it gatom}=5 for Mo, {\it gatom}=6 for W, {\it gatom}=7
 
for Fe (original parameters), {\it gatom}=8 for Fe (modified parameters in erratum).
 
Subtoutine {\bf FS} was coded by Ja,es Elliott in April 2009.
 
   
  +
* [[INTMIN]]: used with [[CHARMM]] keyword to specify minimisation in internal coordinates. This generally appears to be slower than using Cartesian coordinates.
\item {\it GROUND\/}: when combined with keywords {\it NEON\/} or {\it ARGON\/}
 
uses an accurate (Aziz) potential to model the ground state neutral cluster.
 
   
  +
* [[JC]]: Specifies Murrell's two-plus three-body potential.<ref name="murrellm90">
\item {\it GUIDE\/ guidecut}: specifies the RMS force below which the real potential is used
 
  +
<bibtex>
rather than a guiding potential. The systems affected are {\it CPMD\/} and {\it WELCH},
 
  +
@Article{murrellm90,
which are guided by {\it AMBER\/} and {\it TOSI\/}, respectively, and also {\it PACHECO\/},
 
  +
author = {Murrell, J. N. and Mottram, R. E.},
where the Axilrod-Teller contribution is only included when the RMS force falls below
 
  +
title = {potential energy functions for atomic solids},
{\it guidecut\/}. Default {\it guidecut\/}=0.0001.
 
  +
journal = {\molphys},
New guided potentials are {\it ZETT1} and {\it ZETT2} (guided by Morse with $\rho=5$) and
 
  +
year = {1990},
{\it NATB} (guided by {\it GUPTAT}). Parameters for the guiding potential must also be specified in
 
  +
volume = {69},
{\tt data}.
 
  +
pages = {571-585}
  +
}
  +
</bibtex></ref><ref name="murrellr90">
  +
<bibtex>
  +
@Article{murrellr90,
  +
author = {Murrell, J. N. and Rodriguez-Ruiz, J. A.},
  +
title = {potential energy functions for atomic solids 2. potential functions
  +
for diamond-like structures},
  +
journal = {\molphys},
  +
year = {1990},
  +
volume = {71},
  +
pages = {823-834}
  +
}
  +
</bibtex></ref><ref name="alderzijmr91">
  +
<bibtex>
  +
@Article{alderzijmr91,
  +
author = {Alderzi, A. R. and Johnston, R. L. and Murrell, J. N. and
  +
Rodrigez-Ruiz, J. A.},
  +
title = {potential energy functions for atomic solids. 3. fitting phonon
  +
spectra and elastic constants of diamond structures},
  +
journal = {\molphys},
  +
year = {1991},
  +
volume = {73},
  +
pages = {265-282}
  +
}
  +
</bibtex></ref><ref name="eggenjlm92">
  +
<bibtex>
  +
@Article{eggenjlm92,
  +
author = {Eggen, B. E. and Johnston, R. L. and Li, S. and Murrell, J. N.},
  +
journal = {\molphys},
  +
year = {1992},
  +
volume = {76},
  +
pages = {1405}
  +
}
  +
</bibtex></ref><ref name="fengjm93">
  +
<bibtex>
  +
@Article{fengjm93,
  +
author = {Feng, J.-Y. and Johnston, R. L. and Murrell, J. N.},
  +
journal = {\molphys},
  +
year = {1993},
  +
volume = {78},
  +
pages = {1405}
  +
}
  +
</bibtex></ref> A file {\tt JMparams} must exist in the current directory containing the parameters <math>c_0</math>, <math>c_1</math>, <math>\ldots</math>, <math>c_{10}</math>, <math>r_e</math>, <math>D</math>, <math>a_2</math> and <math>a_3</math>. An optional cutoff parameter can also be provided at the end of the <tt>JMparams</tt> file. Subroutines used: '''jmec''', '''jm2c''', '''jm3c'''.
   
  +
* [[JUMPMOVE]] ''np1 np2 int'': specify J-walking type attempts between parallel runs ''np2'' and ''np1'' at intervals of ''int'' steps.
\item{\it GUPTA gatom}: specifies a Gupta potential using parameters from Cleri and Rosato,
 
{\it Phys.~Rev.~B}, {\bf 48}, 22 (1993). {\it gatom=1} for Ni,
 
{\it gatom=2} for Cu,
 
{\it gatom=3} for Rh,
 
{\it gatom=4} for Pd,
 
{\it gatom=5} for Ag,
 
{\it gatom=6} for Ir,
 
{\it gatom=7} for Pt,
 
{\it gatom=8} for Au,
 
{\it gatom=9} for Al,
 
{\it gatom=10} for Pb,
 
{\it gatom=11} for Ti type 1,
 
{\it gatom=12} for Ti type 2,
 
{\it gatom=13} for Zr type 1,
 
{\it gatom=14} for Zr type 2,
 
{\it gatom=15} for Co,
 
{\it gatom=16} for Cd type 1,
 
{\it gatom=17} for Cd type 2,
 
{\it gatom=18} for Zn,
 
{\it gatom=19} for Mg,
 
{\it gatom=20} for V,
 
{\it gatom=21} for Na,
 
{\it gatom=22} for Sr (Wang and Blaisten-Barojas, {\it J.~Chem.~Phys.}, {\bf 115}, 3640 (2001)),
 
{\it gatom=22} for Au as used by Garzon et al.
 
The {\bf Gupta} subroutine was recoded more efficiently by James Elliott in April 2009.
 
   
% \item {\it GTOL\/ gtol}: specifies the convergence criterion for line searches in the
+
* [[LB2]]: specifies the potential <ref name="LB299a">
  +
<bibtex>
% LBFGS routine. Default {\it gtol\/}=0.9.
 
  +
@Article{LB299a,
% This keyword is obsolete, since line searches have been removed.
 
  +
author = {Lynden-Bell, D. and Lynden-Bell, R. M.},
  +
journal = {Proc. Roy. Soc. Lond. A},
  +
year = {1999},
  +
volume = {455},
  +
pages = {475-489}
  +
}
  +
</bibtex></ref><ref name="LB299b">
  +
<bibtex>
  +
@Article{LB299b,
  +
author = {Lynden-Bell, D. and Lynden-Bell, R. M.},
  +
journal = {Proc. Roy. Soc. Lond. A},
  +
year = {1999},
  +
volume = {455},
  +
pages = {3261-3284}
  +
}
  +
</bibtex></ref><ref name="LB204">
  +
<bibtex>
  +
@Article{LB204,
  +
author = {Lynden-Bell, D. and Lynden-Bell, R. M.},
  +
journal = {J. Stat. Phys.},
  +
year = {2004},
  +
volume = {117},
  +
pages = {199-209}
  +
}
  +
</bibtex></ref><math>V = \frac{\epsilon}{2} \sum_{i<j} \left[ \left(\frac{r_{ij}}{\sigma}\right)^2+\left(\frac{\sigma}{r_{ij}}\right)^2 \right]</math>, where <math>\epsilon</math> and <math>\sigma</math> are set to unity.
   
  +
* [[LIGMOVE]] ''ligrotscale ligcartstep'': used with [[AMBER9]] and [[MOVABLEATOMS]]. Specifies ligand only rotation and cartesian perturbations. The ligand is defined by atom index in the file <tt>movableatoms</tt>. Setting ''ligrotscale'' less than 1.0 limits the ammount of rotation possible - this may be required to prevent cold fusion with non-spherical ligands. ''ligcartstep'' defines the maximum size (in angstroms) of the random cartesian perturbations applied to the ligand. Both rotation and cartesian ligand moves are applied AFTER any MD if [[AMBERMDMOVES]] is on to prevent the MD exploding.
% \item {\it HISTOGRAM histmin, histint, histfac, hbins, histfacmul, targetwl, hpercent}: specifies
 
% a basin-sampling calculation\cite{BogdanWC06}
 
% of the energy density of states. The parameters are system-dependent so there are
 
% no default values, but whenever applicable, the recommended values are specified in parentheses.
 
% {\it histmin} is the energy of the lowest bin (usually the energy
 
% of the global minimum of a given potential energy surface),
 
% {\it histint} is the energy difference between bins (roughly $5\%$ of the energy spectrum being sampled),
 
% {\it histfac} is the initial modification factor,
 
% {\it hbins} is the number of bins, which is determined by the energy range to be sampled.
 
% {\it histfacmul} is the power of the power-law that is used to decrease the modification factor at each WL
 
% iteration (any function that allows for smooth convergence of {\it histfac} to unity is acceptable,
 
% using a square root function
 
% {\it histfacmul} = $0.5$ has been found to perform well).
 
% {\it targetwl} is the requested number of WL iterations, which serves as a
 
% convergence criterion for the Wang-Landau sampling.
 
% {\it hpercent} is the histogram flatness criterion, the value by which visits to a given bin are allowed to deviate from the mean
 
% while the histogram is still considered flat.
 
   
  +
* [[LJCOUL]] ''nc'' <math>q'</math> ''f'' <math>T_{\rm swap}</math>: specifies a cluster of Lennard-Jones particles in which the first ''nc'' particles carry identical reduced charges <math>q'</math> in addition to the Lennard-Jones interaction. The parameter ''f'' specifies what fraction of the Monte Carlo steps should be swaps between the positions of a charged and a neutral particle, rather than a conventional step. <math>T_{\rm swap}</math> is the temperature to be used in the acceptance criterion for swap moves, overriding that specified using the [[TEMPERATURE]] keyword. Generally, a lower temperature is more effective at finding the lowest-energy permutation of charges. The default value of <math>T_{\rm swap}</math> is zero. The reduced charge <math>q'</math> is related to the actual charge <math>q</math> by <math>q'=q/(4\pi\epsilon_0\epsilon\sigma)^{1/2}</math>, where <math>\epsilon</math> and <math>\sigma</math> are the Lennard-Jones well depth and length parameter respectively. This way, the reduced energy of two charges is <math>E'=q'^2/r'</math>, where <math>r'=r/\sigma</math> is the reduced distance between the charges.
% \item{\it HISTRESTART}: if present, basin-sampling run reads in the
 
% {\tt lnWeight.his, Distance.his, MinDistance.his,
 
% VisitsTotal.his} statistics and restarts from a WL iteration specified in
 
% {\tt nWL.restart} and {\tt lnModfac.restart}.
 
   
  +
* [[LOCALSAMPLE]] ''abthresh acthresh'': Keyword currently under construction! For three groups of atoms defined in <tt>movableatoms</tt> (A,B,C), a step is quenched when the centre of coordinates of A->B is less than ''abthresh'' AND A->C is less than ''acthresh''. If this condition is broken AFTER the quench, it is automatically rejected.
\item{\it INTMIN}: used with {\it CHARMM} keyword to specify minimisation in internal
 
coordinates. This generally appears to be
 
slower than using Cartesian coordinates.
 
   
  +
=== M-R ===
\item{\it JC}: Specifies Murrell's two- plus three-body
 
  +
* [[MAXBFGS]] ''max'': ''max'' is the largest permitted LBFGS step.
potential.\cite{murrellm90,murrellr90,alderzijmr91,eggenjlm92,fengjm93}
 
A file {\tt JMparams} must
 
exist in the current directory containing the parameters $c_0,\ c_1,\ldots,\ c_{10},\ r_e,\
 
D,\ a_2$ and $a_3$. An optional cutoff parameter can also be provided at the end of the
 
{\tt JMparams} file.
 
Subroutines used: {\bf jmec}, {\bf jm2c}, {\bf jm3c}.
 
   
  +
* [[MAXERISE]] ''maxez'': specifies the largest rise in energy permitted during an LBFGS minimisation, default <math>10^{-10}</math>. Useful for potentials with imprecise derivatives.
\item {\it JUMPMOVE np1 np2 int\/}: specify J-walking type attempts between parallel runs {\it np2\/}
 
and {\it np1\/} at intervals of {\it int\/} steps.
 
   
  +
* [[MAXIT]] ''maxit maxit2'': ''maxit'' and ''maxit2'' are integers specifying the maximum number of iterations allowed in the conjugate gradient quenches. ''maxit'' applies to the 'sloppy' quenches of the basin-hopping run and ''maxit2'' to the final quenches that are used to produce the output in file <tt>lowest</tt>.
\item {\it LB2} specifies the potential\cite{LB299a,LB299b,LB204}
 
\begin{equation}
 
V = \frac{\epsilon}{2} \sum_{i<j} \left[ \left(\frac{r_{ij}}{\sigma}\right)^2+
 
\left(\frac{\sigma}{r_{ij}}\right)^2 \right],
 
\end{equation}
 
where $\epsilon$ and $\sigma$ are set to unity.
 
   
  +
* [[MGGLUE]]: specifies a glue potential for magnesium.
\item {\it LIGMOVE ligrotscale ligcartstep\/}: used with {\it AMBER9\/} and {\it MOVABLEATOMS}. Specifies ligand only rotation and
 
cartesian perturbations. The ligand is defined by atom index in the file 'movableatoms'. Setting {\it ligrotscale} less than 1.0
 
limits the ammount of rotation possible - this may be required to prevent cold fusion with non-spherical ligands. {\it ligcartstep}
 
defines the maximum size (in angstroms) of the random cartesian perturbations applied to the ligand. Both rotation and cartesian ligand
 
moves are applied AFTER any MD if {\it AMBERMDMOVES} is on to prevent the MD exploding.
 
   
  +
* [[MORSE]] ''rho'': specifies a Morse potential with range parameter ''rho''.<ref name="braierbw90">
\item {\it LJCOUL nc $q'$ f $T_{\rm swap}$} specifies a cluster of Lennard-Jones particles in which the first {\it nc}
 
  +
<bibtex>
particles carry identical reduced charges {\it $q'$} in addition to the Lennard-Jones interaction.
 
  +
@Article{braierbw90,
The parameter {\it f} specifies what fraction of the Monte Carlo steps should be swaps between the
 
  +
author = {Braier, P. A. and Berry, R. S. and Wales, D. J.},
positions of a charged and a neutral particle, rather than a conventional step.
 
  +
title = {How the range of pair interactions governs features of
$T_{\rm swap}$ is the temperature to be used in the acceptance criterion for swap moves, overriding
 
  +
multidimensional potentials},
that specified using the {\it TEMPERATURE} keyword. Generally, a lower temperature is more effective
 
  +
journal = {\jcp},
at finding the lowest-energy permutation of charges. The default value of $T_{\rm swap}$ is zero.
 
  +
year = {1990},
The reduced charge $q'$ is related to the actual charge $q$ by $q'=q/(4\pi\epsilon_0\epsilon\sigma)^{1/2}$,
 
  +
volume = {93},
where $\epsilon$ and $\sigma$ are the Lennard-Jones well depth and length parameter respectively.
 
  +
pages = {8745},
This way, the reduced energy of two charges is $E'=q'^2/r'$, where $r'=r/\sigma$ is the reduced distance
 
  +
category = {clusters}
bdetween the charges.
 
  +
}
  +
</bibtex></ref><ref name="doyewb95">
  +
<bibtex>
  +
@Article{doyewb95,
  +
author = {Doye, J. P. K. and Wales, D. J. and Berry, R. S.},
  +
title = {THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURES OF
  +
CLUSTERS},
  +
journal = {\jcp},
  +
year = {1995},
  +
volume = {103},
  +
pages = {4234-4249}
  +
}
  +
</bibtex></ref><ref name="doyew96a">
  +
<bibtex>
  +
@Article{doyew96a,
  +
author = {Doye, J. P. K. and Wales, D. J.},
  +
title = {THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURE AND
  +
STABILITY OF SIMPLE LIQUIDS - FROM CLUSTERS TO BULK, FROM SODIUM TO C-60},
  +
journal = {J. Phys. B},
  +
year = {1996},
  +
volume = {29},
  +
pages = {4859-4894}
  +
}
  +
</bibtex></ref>
   
  +
* [[MPI]]: specifies an MPI parallel job. (only available if the source is compiled with MPI enabled).
\item {\it LOCALSAMPLE abthresh acthresh\/}: Keyword currently under construction! For three groups of atoms defined in movableatoms
 
(A,B,C), a step is quenched when the centre of coordinates of A->B is less than {\it abthresh} AND A->C is less than {\it acthresh}.
 
If this condition is broken AFTER the quench, it is automatically rejected.
 
   
  +
* [[MSORIG]]: specifies a particular tight-binding potential for silicon.
\item {\it MAXBFGS max\/}: {\it max\/} is the largest permitted LBFGS step.
 
   
  +
* [[MSTRANS]]: specifies an alternative tight-binding potential for silicon.
\item {\it MAXERISE maxez\/}: specifies the largest rise in energy permitted during an LBFGS
 
minimisation, default $10^{-10}$. Useful for potentials with imprecise derivatives.
 
   
  +
* [[MULLERBROWN]]: specifies the 2D Muller-Brown potential.
\item {\it MAXIT maxit maxit2\/}: {\it maxit\/} and {\it maxit2\/} are integers specifying the
 
maximum number of iterations allowed in the conjugate gradient quenches. {\it maxit\/} applies
 
to the `sloppy' quenches of the basin-hopping run and {\it maxit2\/} to the final quenches
 
that are used to produce the output in file {\tt lowest}.
 
   
  +
* [[MULTIPLICITY]] ''xmul'': specifies the multiplicity of the electronic state in ''DFTB'' calculations.
\item {\it MGGLUE\/}: specifies a glue potential for magnesium
 
   
  +
* [[NATB]]: specifies the sodium tight-binding potential of Calvo and Spiegelmann. This potential can be guided by also specifying [[GUPTA]] ''21'' in the <tt>data</tt> file.
\item {\it MORSE rho\/}: specifies a Morse potential
 
with range parameter {\it rho\/}.\cite{braierbw90,doyewb95,doyew96a}
 
   
  +
* [[NEON]]: introduces a diatomics-in-molecules calculation for a neutral, cationic or electronically excited neon cluster. See also [[GROUND]], [[PLUS]], [[TWOPLUS]] and [[STAR]].
\item {\it MPI\/}: specifies an MPI parallel job.
 
(only available if the source is compiled with MPI enabled).
 
   
  +
* [[NEWJUMP]] ''prob'': for (serial) 'parallel' runs specifies a jump probability between runs (parallel tempering) of ''prob''. See the [[BHPT]] keyword for a better alternative.
\item {\it MSORIG \/}: specifies a particular tight-binding potential for silicon.
 
   
  +
* [[NEWRESTART]] ''nrelax nhs MD newrestemp'': reseed runs if the energy does not decrease within ''nrelax'' steps. ''nhs'' is the number of hard sphere moves used to produce the new starting configuration. If ''nhs=0'' (the default) then the geometry is changed by reseeding. If ''MD'' is present, then a short AMBER or CHARMM MD run is performed at temperature ''newrestemp'' to generate the new configuration.
\item {\it MSTRANS \/}: specifies an alternative tight-binding potential for silicon.
 
   
  +
* [[NMAX]] ''nmax'': ''nmax'' is an integer that specifies the maximum number of dihedral angles to be twisted in AMBER - '''old implementation'''.
\item {\it MULLERBROWN \/}: specifies the 2D Muller-Brown potential.
 
   
  +
* [[NMIN]] ''nmax'': ''nmin'' is an integer that specifies the minimum number of dihedral angles to be twisted in AMBER - '''old implementation'''.
\item {\it MULTIPLICITY xmul\/}: specifies the multiplicity of the electronic state in {\it DFTB\/}
 
calculations.
 
   
  +
* [[NOCHIRALCHECKS]]: disables checks for inversion of CA atoms and chiral side-chains for ILE and THR.
\item {\it NATB}: specifies the sodium tight-binding potential of Calvo and Spiegelmann.
 
This potential can be guided by also specifying {\it GUPTA 21} in the {\tt data} file.
 
   
  +
* [[NOCISTRANS]] ''minomega'': set on by default with a threshold ''minomega'' of 150 degrees. If an amide bond is deformed to a angle below the specified threshold, the structure is discarded. i.e. with <math>|\omega|<</math>''minomega''. ''minomega'' defaults to 150 degrees. For proline every <math>\omega</math> is allowed. To enable cis-trans isomerisation, the [[CISTRANS]] keyword should be used.
\item {\it NEON\/}: introduces a diatomics-in-molecules calculation for
 
a neutral, cationic or electronically excited neon cluster. See also
 
{\it GROUND\/}, {\it PLUS\/}, {\it TWOPLUS\/} and {\it STAR\/}.
 
   
  +
* [[NOCISTRANSDNA]] ''minomega'': should be specified when working with DNA in AMBER to ensure the correct bonds are checked. As above, the deformation threshold ''minomega'' can be set. It is defaulted to 150 degrees.
\item {\it NEWJUMP prob\/}: for (serial)
 
`parallel' runs specifies a jump probability between runs
 
(parallel tempering) of {\it prob\/}.
 
See the {\it BHPT\/} keyword for a better alternative.
 
   
  +
* [[NOCISTRANSRNA]] ''minomega'': should be specified when working with RNA in AMBER to ensure the correct bonds are checked. As above, the deformation threshold ''minomega'' can be set. It is defaulted to 150 degrees.
\item {\it NEWRESTART nrelax nhs MD newrestemp\/}: reseed runs if the energy does not decrease within {\it nrelax} steps.
 
{\it nhs} is the number of hard sphere moves used to produce the new starting configuration.
 
If {\it nhs=0} (the default) then the geometry is changed by reseeding. If {\it MD} is present, then a short AMBER or CHARMM MD run is performed at temperature {\it newrestemp} to generate the new configuration.
 
   
  +
* [[NOFREEZE]]: don't freeze the core atoms when doing the initial geometry optimisations in a run where [[SEED]] is specified.
\item {\it NMAX nmax\/}: {\it nmax\/} is an integer that specifies the maximum number of dihedral angles
 
to be twisted in {\it AMBER\/}.
 
   
  +
* [[NOPHIPSI]]: used with the [[CHARMM]] keyword to specify twisting of sidechain dihedrals only.
\item {\it NMIN nmax\/}: {\it nmin\/} is an integer that specifies the minimum number of dihedral angles
 
to be twisted in {\it AMBER\/}.
 
   
  +
* [[NORESET]]: by default the configuration point is set to that of the quench minimum in the Markov chain during a basin-hopping simulation. This keyword turns off the resetting so that the geometry varies continuously.
\item {\it NOCHIRALCHECKS\/}: disables checks for inversion of CA atoms and chiral side-chains for ILE and THR.
 
   
  +
* [[NOTE]]: the rest of the line is ignored.
\item {\it NOCISTRANS minomega\/}: set on by default with a threshold {\it minomega} of 150 degrees.
 
If an amide bond is deformed to a angle below the specified threshold, the structure is discarded.
 
i.e.~with $|\omega|<${\it minomega}. {\it minomega\/} defaults to 150 degrees.
 
For proline every $\omega$ is allowed. To enable cis-trans isomerisation, the {\it CISTRANS} keyword should be used.
 
   
  +
* [[ODIHE]]: order parameter - '''requires documentation'''.
\item {\it NOCISTRANSDNA minomega\/}: should be specified when working with DNA in AMBER to ensure the correct bonds are
 
checked. As above, the deformation threshold {\it minomega} can be set. It is defaulted to 150 degrees.
 
   
  +
* [[OEINT]]: interaction energy between 2 peptides will be used as an order parameter - '''requires further documentation'''.
\item {\it NOCISTRANSRNA minomega\/}: should be specified when working with RNA in AMBER to ensure the correct bonds are
 
checked. As above, the deformation threshold {\it minomega} can be set. It is defaulted to 150 degrees.
 
   
  +
* [[ORGYR]]: radius of gyration will be calculated as an order parameter - '''requires further documentation'''.
\item {\it NOFREEZE\/}: don't freeze the core atoms when doing the initial geometry optimisations in
 
a run where {\it SEED\/} is specified.
 
   
  +
* [[OSASA]]: order parameter - '''requires documentation'''.
\item{\it NOPHIPSI}: used with the {\it CHARMM} keyword to specify twisting of
 
sidechain dihedrals only.
 
   
  +
* [[P46]]: specifies a 46-bead three-colour model polypeptide. See also the [[BLN]] keyword, which implements this potential in a more general way and uses unit bond lengths.
\item {\it NORESET\/}: by default the configuration point is set to that of the
 
quench minimum in the Markov chain during a basin-hopping simulation. This
 
keyword turns off the resetting so that the geometry varies continuously.
 
   
  +
* [[PACHECO]]: specifies the intermolecular Pacheco-Ramelho potential for C<math>_{60}</math>. The Axilrod-Teller contribution, specified with the [[AXTELL]] keyword, is included when the RMS force falls below the value entered with ''guidecut''.
\item {\it NOTE \/}: the rest of the line is ignored.
 
   
  +
* [[PAH]]: specifies a polycyclic aromatic hydrocarbon potential.
\item {\it ODIHE \/}: order parameter---requires documentation.
 
   
  +
* [[PARALLEL]] ''npar'': ''npar'' is the number of parallel runs within GMIN.
\item {\it OEINT \/}: interaction energy between 2 peptides will be used as an order parameter---requires
 
further documentation.
 
   
  +
* [[PBGLUE]]: specifies a glue potential for lead.
\item {\it ORGYR \/}: radius of gyration will be calculated as an order parameter---requires
 
further documentation.
 
   
  +
* [[PERIODIC]] ''boxlx boxly boxlz'': specifies periodic boundary conditions for potentials which understand such a directive (such as tight-binding silicon). The three double precision variables are the box lengths. If only one box length is given the others are set to the same value to give a cube.
\item {\it OSASA \/}: order parameter---requires documentation.
 
   
  +
* [[PERMDIST]]: minimise distances between the coordinates in files <tt>coords</tt> and the fixed coordinates in file <tt>finish</tt> with respect to permutational isomerisation. Requires the auxiliary file <tt>perm.allow</tt> to specify permutable atoms, otherwise all atoms are assumed to be permutable. The absence of a <tt>perm.allow</tt> file is considered a mistake for [[CHARMM]] runs. The first line of the <tt>perm.allow</tt> file must contain an integer that specifies the number of groups of interchangeable atoms. The groups then follow, each one introduced by a line with two integers that specify the number of permutable atoms and the number of other pairs of atoms that must swap if the first pair are permuted. The following line contains the numbers of the permutable atoms and then the numbers of the swap pairs, in order. For [[CHARMM]] runs the <tt>perm.allow</tt> file can be generated automatically from the <tt>input.crd</tt> file using the python script [[perm.py]] written by Dr Mey Khalili. Note that it is then ''essential'' to use [[CHARMM]] topology files that include symmetrised definitions of all the relevant amino acids if energies are actually calculated.
\item {\it P46\/}: specifies a 46-bead three-colour model polypeptide.
 
See also the {\it BLN} keyword, which implements this potential in a more
 
general way and uses unit bond lengths.
 
   
  +
* [[PLUS]]: when combined with keywords [[NEON]] or [[ARGON]] uses a diatomics-in-molecules potential for the singly charged cation.
\item {\it PACHECO\/}: specifies the intermolecular Pacheco-Ramelho potential for C$_{60}$.
 
The Axilrod-Teller contribution, specified with the {\it AXTELL\/} keyword, is included
 
when the RMS force falls below the value entered with {\it GUIDECUT\/}.
 
   
  +
* [[PMAX]] ''max'': ''max'' is the maximum probability of dihedral angle twisting for [[ AMBER]] - '''old implementation?'''.
\item{\it PAH}: specifies a polycyclic aromatic hydrocarbon potential.
 
   
  +
* [[PMIN]] ''min'': ''min'' is the minimum probability of dihedral angle twisting for [[AMBER]] - '''old implementation?'''.
\item {\it PARALLEL npar\/}: {\it npar\/} is the number of parallel runs within GMIN.
 
   
  +
* [[POWER]] ''ipow'': ''ipow'' is the initial power for shifts in the old line minimisation routine for conjugate gradient. LBFGS minimisation should be used instead.
\item {\it PBGLUE\/}: specifies a glue potential for lead.
 
   
  +
* [[PROJI]]: turns on projection operator to enforce <math>I</math> point group symmetry in <tt>mylbfgs.f</tt>. The geometry is projected after every proposed step.
\item {\it PERIODIC boxlx boxly boxlz\/}: specifies periodic boundary conditions for
 
potentials which understand such a directive (such as tight-binding silicon). The three
 
double precision variables are the box lengths. If only one box length is given the
 
others are set to the same value to give a cube.
 
   
  +
* [[PROJIH]]: turns on projection operator to enforce <math>I_h</math> point group symmetry in <tt>mylbfgs.f</tt>. The geometry is projected after every proposed step.
\item {\it PERMDIST\/}: minimise distances between
 
the coordinates in files {\tt coords} and the
 
fixed coordinates in file {\tt finish} with respect to permutational isomerisation.
 
Requires the auxiliary file {\tt perm.allow} to specify permutable atoms, otherwise
 
all atoms are assumed to be permutable. The absence of a {\tt perm.allow}
 
file is considered a mistake for {\it CHARMM\/} runs.
 
The first line of the {\tt perm.allow} file must contain an integer
 
that specifies the number of groups of interchangeable atoms.
 
The groups then follow, each one introduced by a line with two integers
 
that specify the number of permutable atoms and the number of other pairs
 
of atoms that must swap if the first pair are permuted.
 
The following line contains the numbers of the permutable atoms and then
 
the numbers of the swap pairs, in order.
 
For {\it CHARMM\/} runs the {\tt perm.allow} file can be generated automatically
 
from the {\tt input.crd} file using the python script {\tt perm.py}
 
written by Dr Mey Khalili.
 
Note that it is then {\it essential} to use {\it CHARMM\/} topology files
 
that include symmetrised definitions of all the relevant amino acids if
 
energies are actually calculated.
 
   
  +
* [[PTMC]] ''histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nenrper hbins'': requests a standard parallel tempering MC run.
\item {\it PLUS\/}: when combined with keywords {\it NEON\/} or {\it ARGON\/}
 
uses a diatomics-in-molecules potential for the singly charged cation.
 
   
  +
This keyword also specifies the energy range for the histogram of quench energies, ''histmin'' to ''histmax'', the energy range for the histogram of instantaneous configurations, ''ptemin'' to ''ptemax'', the temperature range (''pttmin''} and ''pttmax''), the probability of attempting an exchange ''exchprob'', the number of equilibration steps, ''nequil'', the number of parallel tempering MC steps without quenching, ''ptsteps'', the number of bins for the histogram of instantaneous potential energy, ''nenrper'', and the number of bins for the histogram of quench energies, ''hbins''. Should be used together with the [[MPI]] keyword. (This option is only available if the source is compiled with an MPI enabled.)
\item {\it PMAX max\/}: {\it max\/} is the maximum probability of dihedral angle twisting for {\it AMBER\/}.
 
   
  +
* [[PULL]] ''a1 a2 f'': apply a static force to the potential, equivalent to adding the term <math>V_{\rm pull}=-f(z_{a1}-z_{a2})</math>. Here <math>z_{a1}</math> and <math>z_{a2}</math> are the <math>z</math> coordinates for atoms <math>a1</math> and <math>a2</math>, and <math>f</math> specifies the force. This potential is designed to simulate a pulling experiment with static force where a molecule is pulled along the <math>z</math> axis from atoms <math>a1</math> and <math>a2</math>.
\item {\it PMIN min\/}: {\it min\/} is the minimum probability of dihedral angle twisting for {\it AMBER\/}.
 
   
  +
* [[QCUTOFF]] ''qcut'': ''qcut'' is a distance cut-off for Coulomb interactions in ''AMBER''.
\item {\it POWER ipow\/}: {\it ipow\/} is the initial power for shifts in the old line minimisation routine
 
for conjugate gradient. LBFGS minimisation should be used instead.
 
   
  +
* [[QMAX]] ''cgmax'': ''cgmax'' is the tolerance for the RMS force in the final set of quenches that are used to produce the output for file <tt>lowest</tt>. The default is ''cgmax''<math>=10^{-3}</math>, but the appropriate value depends upon the system in question. ''TIGHTCONV'' can be used instead.
\item {\it PROJI\/}: turns on projection operator to enforce $I$ point group symmetry
 
in {\bf mylbfgs.f}. The geometry is projected after every proposed step.
 
   
  +
* [[QUAD]]: '''requires documentation :('''.
\item {\it PROJIH\/}: turns on projection operator to enforce $I_h$ point group symmetry
 
in {\bf mylbfgs.f}. The geometry is projected after every proposed step.
 
   
  +
* [[QUCENTRE]]: sets the centre of coordinates to the origin (0,0,0) before each MC step is taken (so after each quench), but not during the minimisation itself unlike [[CENTRE]].
\item {\it PTMC histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nenrper hbins\/}:
 
requests a standard parallel tempering MC run.
 
This keyword also specifies the energy range for the histogram of quench energies,
 
{\it histmin\/} to {\it histmax\/},
 
the energy range for the histogram of instantaneous configurations, {\it ptemin} to {\it ptemax},
 
the temperature range ({\it pttmin} and {\it pttmax}),
 
the probability of attempting an exchange {\it exchprob}, the
 
number of equilibration steps, {\it nequil},
 
the number of parallel tempering MC steps without quenching, {\it ptsteps},
 
the number of bins for the histogram of instantaneous potential energy, {\it nenrper}, and
 
the number of bins for the histogram of quench energies, {\it hbins}.
 
Should be used together with the {\it MPI\/} keyword. % and {\it BINSTRUCTURES\/} keywords.
 
(This option is only available if the source is compiled with an MPI enabled.)
 
   
  +
* [[RADIUS]] ''radius'': sets the radius of the container that prevents particles evaporating during quenches.
\item {\it PULL a1 a2 f\/}: apply a static force to the potential, equivalent to adding
 
the term $V_{\rm pull}=-f(z_{a1}-z_{a2})$. Here $z_{a1}$ and $z_{a2}$ are the $z$
 
coordinates for atoms $a1$ and $a2$, and $f$ specifies the force.
 
This potential is designed to simulate a pulling experiment with static force where
 
a molecule is pulled along the $z$ axis from atoms $a1$ and $a2$.
 
   
  +
If unset the program calculates an appropriate value based upon the volume per particle for close-packed material and the known pair equilibrium distance for the given potential. The formula employed is:
\item {\it QCUTOFF qcut\/}: {\it qcut\/} is a distance cut-off for Coulomb interactions in {\it AMBER\/}.
 
   
  +
<math>RADIUS = r_e \left[ 1 + \left( \frac{3n}{4\pi \sqrt{2}} \right)^{1/3} \right] </math>
\item {\it QMAX cgmax\/}: {\it cgmax\/} is the tolerance for the
 
RMS force in the final set of quenches that are used to produce
 
the output for file {\tt lowest}. The default is
 
{\it cgmax\/}$=10^{-3}$, but the appropriate value depends upon the system in question.
 
{\it TIGHTCONV} can be used instead.
 
   
  +
where <math>n</math> is the number of atoms and <math>r_e</math> is the pair equilibrium separation. <ref name="kittel76">
\item {\it QUAD\/}: requires documentation.
 
  +
<bibtex>
  +
@Book{kittel76,
  +
author = {Kittel, C.},
  +
title = {Introduction to Solid State Physics},
  +
edition = {third},
  +
year = {1976},
  +
publisher = {Wiley},
  +
address = {New York}
  +
}
  +
</bibtex></ref> The '1' in this formula is to allow some extra space for more open structures.
   
  +
* [[RANDOMSEED]]: specifies that the random number generator should be seeded with system time after each quench, allowing simple parallel use. Currently functional only for the ''CHARMM'' and ''AMBER'' potentials.
\item {\it QUCENTRE\/}: sets the centre of coordinates to the origin (0,0,0) before each MC step is taken (so after each quench), but not during the minimisation itself unlike {\it CENTRE}.
 
   
  +
* [[RANSEED]] ''i'': integer seed for the random number generator.
\item {\it RADIUS radius\/}: sets the radius of the container that prevents particles
 
evaporating during quenches. If unset the program calculates an appropriate value
 
based upon the volume per particle for close-packed material and the known pair
 
equilibrium distance for the given potential. The formula employed is
 
$$ RADIUS=r_e\left[1 + \left(3 n \over 4\pi\sqrt{2}\right)^{1/3}\right], $$
 
where $n$ is the number of atoms and $r_e$ is the pair equilibrium
 
separation.\cite{kittel76} The `1' in this formula is to allow some extra space for
 
more open structures.
 
   
  +
* [[RCUTOFF]] ''rcut'': ''rcut'' is a distance cut-off in ''AMBER'' - '''old implementation'''.
\item {\it RANDOMSEED\/}: specifies that the random number generator should be seeded with system time after each quench, allowing simple parallel use. Currently functional only for the CHARMM and AMBER potentials.
 
   
  +
* [[RESIZE]] ''resize'': all the coordinates are multiplied by ''resize'' after they have been read in, before any other operations are performed. This command is useful for scaling results obtained with one potential for a system with a different pair equilibrium distance.
\item {\it RANSEED i\/}: integer seed for the random number generator.
 
   
  +
* [[RESTART]] ''nrelax nhs'': reseed runs if a step in not accepted within twice ''nrelax'' steps. ''nhs'' is the number of hard sphere moves used to produce the new starting configuration. If ''nhs=0'' (the default) then the geometry is changed by reseeding.
\item {\it RCUTOFF rcut\/}: {\it rcut\/} is a distance cut-off in {\it AMBER\/}.
 
   
  +
* [[RESTORE]] ''dumpfile intEdumpfile'': restore a previous ''GMIN'' run from a ''dumpfile''. The number of basin-hopping steps performed will be the difference between the number requested for the run that produced the ''dumpfile'', minus the number that were completed at the point the ''dumpfile'' was created. This option is not available before version 2.3. If you are using the [[A9INTE]] keyword, you can specify the interaction enthalpy dump file to restore from as a second arguement.
\item {\it RESIZE resize\/}: all the coordinates are multiplied by {\it resize\/} after
 
they have been read in, before any other operations are performed. This command is useful
 
for scaling results obtained with one potential for a system with a different pair
 
equilibrium distance.
 
   
  +
* [[RGCL2]]: specifies a DIM rare gas-Cl<math>_2</math> potential.
\item {\it RESTART\/ nrelax nhs\/}: reseed runs if a step in not accepted
 
within twice {\it nrelax} steps.
 
{\it nhs} is the number of hard sphere moves used to produce the new starting configuration.
 
If {\it nhs=0} (the default) then the geometry is changed by reseeding.
 
   
  +
* [[RINGROTSCALE]] ''factor'': when applying cartesian moves with ''CHARMM'', amino acid rings are moved as rigid units. Setting ''factor'' (default 0.0) between 0.0 and 1.0 will apply a random rotation to these rings during step taking. The suggested value is 0.1 to prevent the regular formation of high energy structures.
\item {\it RESTORE\/ dumpfile intEdumpfile\/}: restore a previous {\tt GMIN} run from a {\it dumpfile}.
 
The number of basin-hopping steps performed will be the difference between the number
 
requested for the run that produced the dumpfile, minus the number that were completed
 
at the point the dumpfile was created. This option is not available before version 2.3.
 
If you are using the {\it A9INTE\/} keyword, you can specify the interaction enthalpy
 
dump file to restore from as a second arguement.
 
   
\item {\it RGCL2\/}: specifies a DIM rare gas-Cl$_2$ potential.
+
* [[RKMIN]]: specifies a Runga-Kutta minimisation scheme. Very inefficient.
   
  +
* [[RMS]] ''rmslimit rmstol rmssave lca best'': used with ''CHARMM'' keyword to specify that the RMSD compared to a reference geometry is calculated.
\item {\it RINGROTSCALE factor\/}: when applying cartesian moves with CHARMM, amino acid rings are moved as rigid units. Setting {\it factor} (default 0.0) between 0.0 and 1.0 will apply a random rotation to these rings during step taking. The suggested value is 0.1 to prevent the regular formation of high energy structures.
 
   
  +
The reference geometry must be given in xyz-format in an additional file <tt>compare</tt>. ''rmssave'' is an integer that specifies the number of lowest energy geometries and RMSD <math>\le</math> ''rmslimit'' to save. Geometries are only saved if their RMSD's are more than ''rmstol'' different. The flag ''lca'' controls whether the all-atom RMSD (''lca''=0) or the <math>C_{\alpha}</math>-RMSD (''lca''=1) is calculated. The flag ''best'' determines which structure is compared to the reference after each quench. ''best''=0 implies the current quench minimum and ''best''=1 implies the current best (lowest energy) minimum. If [[TRACKDATA]] is also specified, the RMSD calculated after each quench is produced in the file <tt>rmsd</tt> in gnuplot readable format.
\item {\it RKMIN\/}: specifies a Runga-Kutta minimisation scheme.
 
Very inefficient.
 
   
  +
=== S-Z ===
\item{\it RMS rmslimit rmstol rmssave lca best}: used with {\it CHARMM} keyword to
 
  +
* [[SAVE]] ''nsave'': ''nsave'' is an integer that specifies the number of lowest energy geometries to save and summarise in the file <tt>lowest</tt>. Arrays are now dynamically allocated, so any positive integer can be specified.
specify that the RMSD compared to a reference geometry is calculated. The reference geometry must
 
be given in xyz-format in an additional file {\tt compare}. {\it rmssave} is an integer
 
that specifies the number of lowest energy geometries and RMSD $\le$ {\it rmslimit}
 
to save. Geometries are only saved if their RMSD's are more than {\it rmstol}
 
different. The flag {\it lca} controls whether the all-atom RMSD ({\it lca}=0) or the $C_{\alpha}$-RMSD
 
({\it lca}=1) is calculated. The flag {\it best} determines which structure is compared to the reference
 
after each quench. {\it best}=0 implies the current quench minimum and {\it best}=1 implies the current best (lowest energy) minimum. If {\it TRACKDATA} is also specified, the RMSD calculated after each quench is produced in the file `rmsd' in gnuplot readable format.
 
   
\item {\it SAVE nsave\/}: {\it nsave\/} is an integer that specifies the number of lowest
+
* [[SAVEINTE]] ''nsaveinte''}: ''nsaveinte'' is an integer that specifies the number of lowest interaction enthalpy geometries to save and summarise in the file <tt>intelowest</tt>. See [[A9INTE]].
energy geometries to save and summarise in the file {\tt lowest}.
 
Arrays are now dynamically allocated, so any positive integer can be specified.
 
   
  +
* [[SETCENTRE]] ''x y z'': Sets the centre of mass/coordinates (before the initial quench) to (''x,y,z''). For example, [[SETCENTRE]] ''0.0 0.0 0.0'' would translate the centre of mass to the origin.
\item {\it SAVEINTE nsaveinte\/}: {\it nsaveinte\/} is an integer that specifies the number of lowest
 
interaction enthalpy geometries to save and summarise in the file {\tt intelowest}. See {\it A9INTE\/}.
 
   
  +
* [[SC]] ''nn mm sig sceps scc'': specifies a Sutton-Chen potential <ref name="suttonc90">
\item {\it SETCENTRE x y z\/}: Sets the centre of mass/coordinates (before the initial quench) to ({\it x,y,z\/}). For example, {\it SETCENTRE 0.0 0.0 0.0\/}
 
  +
<bibtex>
would translate the centre of mass to the origin.
 
  +
@Article{suttonc90,
  +
author = {Sutton, A. P. and Chen, J.},
  +
title = {long-range Finnis-Sinclair potentials},
  +
journal = {Phil. Mag. Lett.},
  +
year = {1990},
  +
volume = {61},
  +
pages = {139}
  +
}
  +
</bibtex></ref> with parameters <math>n=</math>''nn'', <math>m=</math>''mm'', <math>a</math>=''sig'', <math>\epsilon=</math>''sceps'' and <math>c=</math>''scc''.
   
  +
* [[SEED]] ''nsstop'': if the ''SEED'' keyword appears then the program looks for a file <tt>seed</tt> containing coordinates, which are used to 'seed' the new run. The number of coordinates given in this file should be no more than one less than the number given in <tt>coords</tt>. The specified coordinates are frozen from the first step until step ''nsstop''.
\item {\it SC nn mm sig sceps scc\/}: specifies a Sutton-Chen potential\cite{suttonc90} with
 
parameters $n=${\it nn\/}, $m=${\it mm\/}, $a$={\it sig\/}, $\epsilon=${\it sceps\/} and
 
$c=${\it scc\/}.
 
   
  +
* [[SHIFTCUT]]: specifies a shifted-truncated potential for bulk binary Lennard-Jones.
\item {\it SEED nsstop\/}: if the {\it SEED\/} keyword appears then the program
 
looks for a file {\tt seed} containing coordinates, which are used to `seed' the new run.
 
The number of coordinates given in this file should be no more than one less than the number
 
given in {\tt coords}. The specified coordinates are frozen from the first step until
 
step {\it nsstop\/}.
 
   
  +
* [[SIDESTEP]] ''smax'': specifies the maximum step in Cartesian coordinates for side-chains
\item {\it SHIFTCUT\/}: specifies a shifted-truncated potential for bulk binary Lennard-Jones.
 
  +
in ''AMBER'' - '''old implementation'''.
   
  +
* [[SLOPPYCONV]] ''bgmax'': specifies a basin-hopping run (as opposed to standard MC on the untransformed surface). ''bgmax'' is the convergence criterion for the RMS force in the basin-hopping quenches. If this criterion is too strict then the run time will be greatly increased. If it is too sloppy then the performance of the algorithm is impaired. Different values are needed for different potentials. [[BASIN]] can be used instead.
\item {\it SIDESTEP smax\/}: specifies the maximum step in Cartesian coordinates for side-chains
 
in {\it AMBER\/}.
 
   
  +
* [[SORT]]: for pairwise potentials the atoms can be sorted from most to least strongly bound. The ''SORT'' keyword enables this sorting for the coordinates printed in file <tt>lowest</tt>. This can be useful for seeding subsequent runs by removing the most weakly bound atoms. This sort is not set by default and is meaningless if the pair energies are not computed.
\item {\it SLOPPYCONV bgmax\/}: specifies a basin-hopping run (as opposed to standard MC
 
on the untransformed surface). {\it bgmax\/} is the convergence criterion
 
for the RMS force in the basin-hopping
 
quenches. If this criterion is too strict then the run time will be greatly increased.
 
If it is too sloppy then the performance of the algorithm is impaired. Different values
 
are needed for different potentials. {\it BASIN} can be used instead.
 
   
  +
* [[STAR]]: specifies an excited state calculation for Ar<math>^*_n</math> or Ne<math>^*_n</math> for a diatomics-in-molecules potential when used with [[NEON]] or [[ARGON]].
\item {\it SORT}: for pairwise potentials the atoms can be sorted from most to least
 
strongly bound. The {\it SORT} keyword enables this sorting for the coordinates printed
 
in file {\tt lowest}. This can be useful for seeding subsequent runs by removing the
 
most weakly bound atoms. This sort is not set by default and is meaningless if the
 
pair energies are not computed.
 
   
  +
* [[STEP]] ''step astep ostep block'': specifies the maximum step sizes.
\item {\it STAR}: specifies an excited state calculation for Ar$^*_n$ or Ne$^*_n$ for
 
a diatomics-in-molecules potential when used with {\it NEON\/} or {\it ARGON\/}.
 
   
  +
''step'' is for the maximum change of any Cartesian coordinate and ''astep'' specifies a tolerance on the binding energy of individual atoms (if available, i.e. for Morse and LJ) below which an angular step is taken for that atom. ''ostep'' is the maximum displacement of an axis-angle coordinate for a rigid body system and ''block'' (an integer) is the block size for which separate translational and orientational displacements will be made for rigid bodies. Omitting ''block'' or using a value of zero results in translational and orientational steps being taken simultaneously for rigid bodies. The default values for ''step'', ''astep'' and ''ostep'' are all 0.3 and the default value of ''nblock'' is zero.
\item {\it STEP step astep ostep block\/}: specifies the maximum step sizes. {\it step\/} is
 
for the maximum change of any Cartesian coordinate and {\it astep\/} specifies a tolerance
 
on the binding energy of individual atoms (if available, i.e.~for Morse and LJ) below
 
which an angular step is taken for that atom. See the following section for more details.
 
{\it ostep\/} is the maximum displacement of an axis-angle coordinate for a rigid body system
 
and {\it block\/} (an integer) is the block size for which separate translational and orientational
 
displacements will be made for rigid bodies. Omitting {\it block\/} or using a value of zero results in
 
translational and orientational steps being taken simultaneously
 
for rigid bodies. The default values for {\it step\/},
 
{\it astep\/} and {\it ostep\/} are all 0.3 and the default value of {\it nblock\/} is zero.
 
   
\item {\it STEEREDMIN smink sminkinc smindiststart smindistfinish sminatoma sminatomb\/}: specified steered
+
* [[STEEREDMIN]] ''smink sminkinc smindiststart smindistfinish sminatoma sminatomb'': specified steered minimisation should be performed (must be used with [[AMBER9]]).
minimisation should be performed (must be used with {\it AMBER9}). For a protein/ligand system, this adds a translation
 
to the MC move. The vector between the centre of coordinates of groups A and B (as defined in the file movableatoms)
 
is calculated and set to {\it smindiststart}. During the following minimisation, a restoring force is applied to
 
the ligand. The harmonic force constant is initially zero, and rises by {\it sminkinc} every LBFGS step up to a
 
maximum of {\it smink}. The force is applied until the A->B distance is less than {\it smindistfinish}.
 
   
  +
For a protein/ligand system, this adds a translation to the MC move. The vector between the centre of coordinates of groups A and B (as defined in the file <tt>movableatoms</tt>) is calculated and set to ''smindiststart''. During the following minimisation, a restoring force is applied to the ligand. The harmonic force constant is initially zero, and rises by ''sminkinc'' every LBFGS step up to a maximum of ''smink''. The force is applied until the A->B distance is less than ''smindistfinish''.
\item {\it STEPS mcsteps tfac\/}: determines the length of the
 
basin-hopping run through the integer {\it mcsteps\/} and the annealing protocol through
 
the real variable {\it tfac\/}. The temperature is multiplied by {\it tfac\/}
 
after every step in each run.
 
   
  +
* [[STEPS]] ''mcsteps tfac'': determines the length of the basin-hopping run through the integer ''mcsteps'' and the annealing protocol through the real variable ''tfac''. The temperature is multiplied by ''tfac'' after every step in each run.
\item {\it STICKY nrbsites, sigma\/}: specifies a `sticky patch' potential with {\it nrbsites}
 
sites in the rigid body reference and a value of {\it sigma} for the $\sigma$ parameter.
 
   
  +
* [[STICKY]] ''nrbsites, sigma'': specifies a 'sticky patch' potential with ''nrbsites'' sites in the rigid body reference and a value of ''sigma'' for the <math>\sigma</math> parameter.
\item {\it STOCK mu lambda}: specifies a Stockmeyer potential with parameters
 
$\mu$ and $\lambda$, respectively.
 
   
  +
* [[STOCK]] ''mu lambda'': specifies a Stockmeyer potential with parameters <math>\mu</math> and <math>\lambda</math>, respectively.
\item {\it STRAND}: specifies a system of $\beta$ strands coded using the rigid body formalism.
 
   
  +
* [[STRAND]]: specifies a system of <math>\beta</math> strands coded using the rigid body formalism.
\item {\it SW\/}: specifies the Stillinger-Weber Si potential.
 
   
  +
* [[SW]]: specifies the Stillinger-Weber Si potential.
\item {\it SYMMETRISE int tol1 tol2 tol3 tol4 tol5 qmax mdiff d}: specifies that the symmetrisation
 
routine should be called every {\it int} steps. The five {\it tol} parameters are tolerances
 
for various parts of the routine:
 
{\it tol1} is used in {\bf ptgrp.f} in defining orbits;
 
{\it tol2} is the distance tolerance used in {\bf ptgrp.f} to define point group symmetry operations;
 
{\it tol3} is the maximum relative difference in principal moments of inertia used to
 
diagnose point groups with degenerate irreducible representations in {\bf ptgrp.f};
 
{\it tol4} is the distance cutoff used to determine if a symmetry element has been lost in {\bf symmetry.f}.
 
Since we are dealing with approximate symmetries, this parameter may be larger than {\it tol2}.
 
It is compared to the largest atomic displacement divided by the corresponding radius
 
for the closest permutation.
 
{\it tol4} is also used to test whether atoms lie on a given symmetry element, and in testing
 
whether orbits generated from `floaters' are actually contained in the core.
 
{\it tol5} is generally to check for atom clashes in {\bf symmetry.f}, including analysis of
 
missing sites in orbits, as well as overlap between orbits generated from `floaters' and
 
previous core or new orbit sites.
 
{\it qmax} is the maximum number of quenches allowed for each call to {\bf symmetry.f}.
 
{\it mdiff} is used to test whether a generated symmetry operation is new. If any of the nine
 
components of the corresponding $3\times3$ matrix differs by more than {\it mdiff} from an
 
existing matrix then the operations are considered to be different.
 
{\it d} is the exponential factor used in constructing a centre of mass that is biased towards
 
core atoms. The contribution of each atom is weighted by $\exp(-dx(i))$, where $x$ is the
 
centre of mass distance of atom $i$ on the previous cycle.
 
   
  +
* [[SYMMETRISE]] ''int tol1 tol2 tol3 tol4 tol5 qmax mdiff d'': specifies that the symmetrisation routine should be called every ''int'' steps.
\item {\it TABOO nlist\/}: specifies a taboo list of the {\it nlist\/} lowest minima should be maintained.
 
   
  +
The five ''tol'' parameters are tolerances for various parts of the routine: ''tol1'' is used in <tt>ptgrp.f</tt> in defining orbits; ''tol2'' is the distance tolerance used in <tt>ptgrp.f</tt> to define point group symmetry operations; ''tol3'' is the maximum relative difference in principal moments of inertia used to diagnose point groups with degenerate irreducible representations in <tt>ptgrp.f</tt>; ''tol4'' is the distance cutoff used to determine if a symmetry element has been lost in <tt>symmetry.f</tt>. Since we are dealing with approximate symmetries, this parameter may be larger than ''tol2''. It is compared to the largest atomic displacement divided by the corresponding radius for the closest permutation. ''tol4'' is also used to test whether atoms lie on a given symmetry element, and in testing whether orbits generated from 'floaters' are actually contained in the core. ''tol5'' is generally to check for atom clashes in <tt>symmetry.f</tt>, including analysis of missing sites in orbits, as well as overlap between orbits generated from 'floaters' and previous core or new orbit sites. ''qmax'' is the maximum number of quenches allowed for each call to <tt>symmetry.f</tt>. ''mdiff'' is used to test whether a generated symmetry operation is new. If any of the nine components of the corresponding <math>3\times3</math> matrix differs by more than ''mdiff'' from an existing matrix then the operations are considered to be different. ''d'' is the exponential factor used in constructing a centre of mass that is biased towards core atoms. The contribution of each atom is weighted by <math>\exp(-dx(i))</math>, where <math>x</math> is the centre of mass distance of atom <math>i</math> on the previous cycle.
\item {\it TARGET target1 target2 $\cdots$\/}: specifies any number of target energies.
 
The current run stops in an orderly
 
fashion if the current quench energy is within {\it econv\/} of any target (see {\it EDIFF\/}).
 
   
  +
* [[TABOO]] ''nlist'': specifies a taboo list of the ''nlist'' lowest minima should be maintained.
\item {\it TEMPERATURE temp\/}: defines the temperature, {\it temp\/}, at which the
 
MC runs are conducted. Different values can be specified for serial `parallel' runs if
 
{\it PARALLEL} is set.
 
For true parallel basin-hopping use the {\it BHPT\/} keyword and omit {\it TEMPERATURE\/}.
 
   
  +
* [[TARGET]] ''target1 target2'' <math>\cdots</math>: specifies any number of target energies. The current run stops in an orderly fashion if the current quench energy is within ''econv'' of any target (see ''EDIFF'').
\item {\it TETHER hdistconstraint hwindows ExtrapolationPercent lnHarmFreq}: requests a calculation of the vibrational density of
 
states for a given minimum. {\it hdistconstraint} is the minimised average radius of the basin of attraction to which the minimum
 
belongs, {\it hwindows} is the number of potential energy windows into which a WL simulation is split. {\it
 
ExtrapolationPercent} is the percentage of the whole potential energy spectrum, for which the density of states is estimated from
 
the harmonic approximation and not sampled. {\it lnHarmFreq} the log product of positive Hessian
 
eigenvalues.
 
   
  +
* [[TEMPERATURE]] ''temp'': defines the temperature, ''temp'', at which the
\item{\it THOMSON q\/}: specify the Thomson problem for unit charges on a sphere.
 
  +
MC runs are conducted. Different values can be specified for serial 'parallel' runs if [[PARALLEL]] is set. For true parallel basin-hopping use the [[BHPT]] keyword and omit [[TEMPERATURE]].
If {\it q\/} is present it is taken to be the charge on one particle, which can
 
therefore be different from all the other unit charges and is read as a real number.
 
   
  +
* [[TETHER]] ''hdistconstraint hwindows ExtrapolationPercent lnHarmFreq'': requests a calculation of the vibrational density of states for a given minimum. ''hdistconstraint'' is the minimised average radius of the basin of attraction to which the minimum belongs, ''hwindows'' is the number of potential energy windows into which a WL simulation is split. ''ExtrapolationPercent'' is the percentage of the whole potential energy spectrum, for which the density of states is estimated from the harmonic approximation and not sampled. ''lnHarmFreq'' the log product of positive Hessian eigenvalues.
% Doesn't appear to be coded?
 
% \item {\it THRESHOLD \/}: specifies threshold acceptance of steps. The change in potential energy must be
 
% less than the value of the {\it TEMPERATURE\/} variable for a step to be accepted.
 
   
  +
* [[THOMSON]] ''q'': specify the Thomson problem for unit charges on a sphere. If ''q'' is present it is taken to be the charge on one particle, which can therefore be different from all the other unit charges and is read as a real number.
\item {\it TIGHTCONV cgmax\/}: cgmax is the tolerance for the
 
RMS force in the final set of quenches that are used to produce
 
the output for file {\tt lowest}. The default is
 
{\it cgmax\/}$=10^{-3}$, but the appropriate values depend upon the system in question.
 
{\it QMAX} can be used instead.
 
   
  +
* [[TIGHTCONV]] ''cgmax'': ''cgmax'' is the tolerance for the RMS force in the final set of quenches that are used to produce the output for file <tt>lowest</tt>. The default is ''cgmax''<math>=10^{-3}</math>, but the appropriate values depend upon the system in question. [[QMAX]] can be used instead.
\item {\it TIP n\/}: specifies a TIP{\it n\/}P intermolecular potential for rigid body water molecules.
 
$\ \le n \le 5$.
 
   
  +
* [[TIP]] ''n'': specifies a TIP''n''P intermolecular potential for rigid body water molecules. <math>\ \le n \le 5</math>.
\item {\it TOLBRENT tolb\/}: parameter for {\it DBRENT\/} minimisation.
 
Inefficient compared to LBFGS.
 
   
  +
* [[TOLBRENT]] ''tolb'': parameter for [[DBRENT]] minimisation. Inefficient compared to LBFGS.
\item{\it TOMEGA}: used with the {\it CHARMM} keyword to specify that peptide bonds will be twisted along with all other dihedrals.
 
   
  +
* [[TOMEGA]]: used with the [[CHARMM]] keyword to specify that peptide bonds will be twisted along with all other dihedrals.
% \item {\it TN\/}: specifies a truncated Newton minimisation scheme.
 
% Inefficient compared to LBFGS.
 
   
\item {\it TOSI app amm apm rho\/}: specifies the Tosi-Fumi potential\cite{tosif64}
+
* [[TOSI]] ''app amm apm rho'': specifies the Tosi-Fumi potential <ref name="tosif64">
  +
<bibtex>
with parameters $A_{++}$, $A_{--}$, $A_{+-}$ and $\rho$.
 
  +
@Article{tosif64,
  +
author = {Tosi, M. P. and Fumi, F. G.},
  +
journal = {J. Phys. Chem. Solids},
  +
year = {1964},
  +
volume = {25},
  +
pages = {45}
  +
}
  +
</bibtex></ref> with parameters <math>A_{++}</math>, <math>A_{--}</math>, <math>A_{+-}</math> and <math>\rho</math>.
   
  +
* [[TRACKDATA]]: produces <tt>energy.dat</tt> and <tt>markov.dat</tt> containing the quench number and associated energy and markov energy in two columns and <tt>best.dat</tt>, containing the current quench number and the current lowest total energy. If ''RMS'' is also specified, a file called <tt>rmsd.dat</tt> is produced containing the RMSD from a reference structure. See ''RMS'' for more information. This allows plotting with gnuplot to monitor convergence of multiple runs. If [[A9INTE]] is also specified, two additional output files are produced, <tt>intE.dat</tt> containing the quench number and associated interaction enthalpy, and <tt>bestintE.dat</tt> containing the quench number and current lowest interaction enthalpy. This keyword does not yet function for MPI runs.
\item {\it TRACKDATA}: produces `energy.dat' and `markov.dat' containing the quench number and
 
associated energy and markov energy in two columns and `best.dat', containing the current quench number and the current lowest
 
total energy. If {\it RMS\/} is also specified, a file called `rmsd.dat' is produced containing the RMSD from a reference structure.
 
See {\it RMS\/} for more information. This allows plotting with gnuplot to monitor convergence of multiple runs.
 
If {\it A9INTE} is also specified, two additional output files are produced, `intE.dat' containing the quench number and associated interaction
 
enthalpy, and `bestintE.dat' containing the quench number and current lowest interaction enthalpy. This keyword does not yet function for MPI runs.
 
   
\item {\it TSALLIS q\/}: specifies that steps are accepted/rejected using Tsallis statistics with the
+
* [[TSALLIS]] ''q'': specifies that steps are accepted/rejected using Tsallis statistics with the given value of ''q'', rather than the usual Boltzmann condition.
given value of {\it q\/}, rather than the usual Boltzmann condition.
 
   
\item {\it TWOPLUS\/}: when combined with keywords {\it NEON\/} or {\it ARGON\/}
+
* [[TWOPLUS]]: when combined with keywords [[NEON]] or [[ARGON]] uses a diatomics-in-molecules potential for the doubly charged cation.
uses a diatomics-in-molecules potential for the doubly charged cation.
 
   
\item {\it UACHIRAL\/}: MUST be included when using ff03ua, the AMBER united atom forcefield unless you have disabled the checks for inverted chiral carbons
+
* [[UACHIRAL]]: '''MUST''' be included when using ''ff03ua'', the ''AMBER'' united atom forcefield unless you have disabled the checks for inverted chiral carbons with [[NOCHIRALCHECKS]]. [[UACHIRAL]] ensures the correct impropers are used to define sidechain chirality when HB hydrogen is missing.
with {\it NOCHIRALCHECKS\/}. {\it UACHIRAL\/} ensures the correct impropers are used to define sidechain chirality when HB hydrogen is missing.
 
   
  +
* [[UNFREEZE]] ''n1 n2'' <math>\ldots</math>: unfreeze the coordinates of atoms ''n1, n2,''<math>\ldots</math>. Only functions in conjunction with [[FREEZEALL]].
\item {\it UPDATES nup\/}: {\it UPDATES\/} is the number of previous steps saved in the LBFGS routine,
 
  +
  +
* [[UNFREEZERES]] ''n1 n2'' <math>\ldots</math>: unfreeze the coordinates of all atoms in residues ''n1, n2,''<math>\ldots</math>. Only functions in conjunction with [[FREEZEALL]].
  +
  +
* [[UPDATES]] ''nup'': ''nup'' is the number of previous steps saved in the LBFGS routine,
 
default 4.
 
default 4.
   
\item {\it VGW ljsigma ljepsilon taumaxsg taumaxfg}: Specifies use of VGW quantum quenching in place of
+
* [[VGW]] ''ljsigma ljepsilon taumaxsg taumaxfg'': Specifies use of VGW quantum quenching in place of classical minimization routines such as LBFGS.
classical minimization routines such as LBFGS. {\it ljsigma} and {\it ljepsilon} are the corresponding Lennard-Jones
 
parameters that must be specified, and taumaxsg and taumaxfg are the maximum value of ``imaginary'' time $\tau$ (inverse tempertaure) for the propagation.
 
The former pertains to the faster ``single-particle'' SP-VGW used for quenching during the MC runs, and the latter for the more accurate
 
``fully-coupled'' VGW used for the final quenching (analogous to the tight convergence of the LBFGS). A $\tau$ of at least
 
2.5 is recommended for the SP-VGW and 5.0 for the FC-VGW. A file {\it vgwdata} containing the masses (in a.m.u.) of all particles, in order of the location
 
of their {\it xyz} coordinates in ``{\it coords}'' must be present (e.g. for a 38 atom Ne cluster, {\it vgwdata} will have 38 lines of ``{\it 20}''). Different
 
masses are permitted, though the current version allows for only one set of LJ parameters.
 
   
  +
''ljsigma'' and ''ljepsilon'' are the corresponding Lennard-Jones parameters that must be specified, and ''taumaxsg'' and ''taumaxfg'' are the maximum value of 'imaginary' time <math>\tau</math> (inverse tempertaure) for the propagation. The former pertains to the faster 'single-particle' SP-VGW used for quenching during the MC runs, and the latter for the more accurate 'fully-coupled' VGW used for the final quenching (analogous to the tight convergence of the LBFGS). A <math>\tau</math> of at least 2.5 is recommended for the SP-VGW and 5.0 for the FC-VGW. A file <tt>vgwdata</tt> containing the masses (in a.m.u.) of all particles, in order of the location of their '''xyz''' coordinates in <tt>coords</tt> must be present (e.g. for a 38 atom Ne cluster, <tt>vgwdata</tt> will have 38 lines of '20'). Different masses are permitted, though the current version allows for only one set of LJ parameters.
\item{\it VGWCPS on magnitude}: Specifies use of contraining potential for SP-VGW (sloppy convergence), as clusters expand during quantum quenching
 
with decreasing mass. 1 or 0 for {\it on} corresponds to on/off,
 
and magnitude should range from 1 to 1000, with 1 having minimal effect, 1000 being highly constrained. Default value is ``on'', with magnitude 1.
 
   
  +
* [[VGWCPS]] ''on magnitude''}: Specifies use of contraining potential for SP-VGW (sloppy convergence), as clusters expand during quantum quenching with decreasing mass. 1 or 0 for ''on'' corresponds to on/off, and magnitude should range from 1 to 1000, with 1 having minimal effect, 1000 being highly constrained. Default value is ''on'', with magnitude 1.
\item{\it VGWCPF on magnitude}: Same as VGWCPS but for FC-VGW, used for the final, full quenching (tight convergence).
 
   
\item{\it VGWTOL magnitude}: Absolute tolerance parameter for differential equation solver used for VGW quenching. Default value is 0.0001.
+
* [[VGWCPF]] ''on magnitude'': Same as [[VGWCPS]] but for FC-VGW, used for the final, full quenching (tight convergence).
  +
For highly quantum or ``stiff'' systems this may need to be increased, while it may be decreased for ``softer'' or less quantum systems to enhance
 
  +
* [[VGWTOL]] ''magnitude'': Absolute tolerance parameter for differential equation solver used for VGW quenching. Default value is 0.0001. For highly quantum or 'stiff' systems this may need to be increased, while it may be decreased for 'softer' or less quantum systems to enhance speed.
speed.
 
 
 
\item {\it VISITPROP}: if specified the Wang-Landau convergence is governed by proportionality of visits to the current value of
+
* [[VISITPROP]]: if specified the Wang-Landau convergence is governed by proportionality of visits to the current value of the modification factor, and not the histogram flatness criterion <ref name="ZhouB03">
  +
<bibtex>
the modification factor, and not the histogram flatness criterion \cite{ZhouB03}.
 
  +
@Article{ZhouB03,
  +
author = {C. Zhou and R.N. Bhatt},
  +
title = {Understanding and Improving the Wang-Landau Algorithm},
  +
journal = {\pre},
  +
year = {2005},
  +
volume = {72},
  +
pages = {025701}
  +
}
  +
</bibtex></ref>.
   
\item {\it WELCH $A_{++}\ A_{--}\ A_{+-}\ \rho\ Q_+\ Q_-\ \alpha_+\ \alpha_-$\/}: specifies a Welch binary
+
* [[WELCH]] <math>A_{++}\ A_{--}\ A_{+-}\ \rho\ Q_+\ Q_-\ \alpha_+\ \alpha_-</math>: specifies a Welch binary salt potential with the parameters indicated.
salt potential with the parameters indicated.
 
   
\item {\it ZETT1\/} and {\it ZETT2\/}: specify the Zetterling potentials.
+
* [[ZETT1]] and [[ZETT2]]: specify the Zetterling potentials.
\end{itemize}
 
   
 
= Some recognised systems =
 
= Some recognised systems =

Latest revision as of 11:20, 20 September 2013

Error creating thumbnail: Unable to save thumbnail to destination
GMIN in all its glory

GMIN page - very much under construction!

Introduction

GMIN is a program that attempts to find the global potential energy minimum for a collection of atoms or molecules using the `basin-hopping' algorithm described by Wales and Doye. [1] A constant temperature Monte Carlo (MC) run is performed on the transformed potential energy surface (PES), and the configuration point may either be reset to the latest minimum in the chain or vary freely. The program knows many different empirical potentials, and it is straightforward to add new systems. From version 2.2 basin-sampling thermodynamics has been added, and from version 2.3 parallel tempering basin-sampling and basin-hopping have been implemented with MPI.

To start a calculation you need a file called data in the current directory, along with a file called coords containing the initial coordinates, which can be random. If the SEED keyword is present in data you also need a file called seed containing the seed coordinates. Most output is written to stdout, although the file lowest is always created at the end of the run, containing the energies and geometries of the lowest few configurations found in the given run. The geometries are saved in XMakemol xyz format.

Source code

You can download the latest version of the source code for all our programs under the GPL here. This tarball does not contain the CHARMM and AMBER source. If you have a license and want to use them with GMIN, please contact Professor Wales.

Keywords

I'm still deciding how to break the keywords up - it is currently done alphabetically - but it might make sense to do it by usage i.e. keywords that work only with AMBER or CHARMM in different sections:

AMBER keywords (old implementation)

AMBER9 keywords

CHARMM keywords

AMH keywords

Feedback needed!--Csw34 09:51, 27 July 2010 (UTC)

The data file

Input is keyword driven with sensible defaults in most cases. Free format may be used within each line. Blank lines are ignored.

The following keywords are recognised, where n and x are integer and real data, respectively.

#-F

  • 2D: enforce two-dimensional `flatland'.
  • A9DIHE resid atom1 atom2 atom3 atom4 step: specifies LeaP should be used to take a step in the dihedral for the atoms specified. It is limited to dihedrals within a residue. This keyword is not yet fully functional.
  • A9INTE: specifies that after each quench that does not lead to an inversion of chirality, isomerisation of a peptide bond or cold fusion - the interaction enthalpy between a specified residue and the rest of the system should be calculated using the external script 'AMBGMINintE.sh', and read back into GMIN.

This is intended for use with protein/ligand systems where you are searching for low energy docked structures. As the total energy does not fully correlate with the protein/ligand interaction enthalpy, it is often useful to retain not only the lowest SAVE total energy structures, but also the lowest SAVEINTE interaction enthalpy structures. To use this keyword, the `AMBGMINintE.sh' script (contained in the SVN repository in the SCRIPTS directory) must be present in the GMIN working directory. You should ensure that you have edited it to match the residue numbering of your system. You also need a full AMBER9+ installation with access to the `sander' executable. When using this keyword, an interaction enthalpy dump file is produced very DUMPINT steps, and at the end of the run, structural output files are produced for the SAVEINTE lowest interaction enthalpy geometries. After each quench, the structure with the current lowest interaction energy is dumped in pdb and rst format prefixed with `bestint.' to allow monitoring.

  • ACCEPTRATIO accrat: accrat is the required acceptance ratio for the MC exploration of the transformed surface. For fixed temperature runs (the default) the maximum step size is adjusted to try and meet the requested value of accrat for a fixed maximum step size the temperature is adjusted instead. The default value of accrat is a half.
  • ACKLAND id: specifies an Ackland embedded atom metal potential coded by Dr Mihai-Cosmin Marinica. id specifies the particular metal: 1 is ?, 2 is ?, 3 is ?, 4 is ?, 5 is iron, 6 is a different iron, 7 is tungsten. Positive values for id specify periodic boundary conditions, where box lengths must be specified by the PERIODIC keyword. Negative values for id specify a cluster calculation. A CUTOFF value can also be used for clusters.
  • ALGLUE: specifies a glue potential for aluminium.
  • AMBER: specifies the AMBER force field (WARNING: old implementation). See also DIELEC.
  • AMBER9 inpcrd inpcrdformat: specifies a calculation with the interfaced version of the Amber 9 program package.

From this package the Amber force fields are being used, with small modifications (e.g. smooth cut-offs). Starting coordinates are read from inpcrd instead (default coords.inpcrd), in Amber inpcrd file format specified by the second optional argument inpcrdformat. If the second argument is missing, it is assumed that inpcrd contains only three columns with the xyz coordinates of all atoms, in the same order as in the topology file. To start a run with this interface, several auxiliary files are required in the same directory: input coordinate file coords.inpcrd, parameter topology file coords.prmtop, input file to Amber containing force field specifications min.in, and, if desired, a coordinate file different from coords.inpcrd containing starting coordinates. To turn on smooth cutoffs for the Generalised Born force fields, the keyword ifswitch=1 has to be used in the \&cntrl namelist block of min.in. When using the AMBER9 keyword, any calculated second derivatives will be numerical. If one wants analytical second derivatives, the NAB keyword should be used instead, with the same syntax. Additional keywords for the AMBER 9 runs are DUMPSTRUCTURES, AMBERMDSTEPS, LIGMOVE (0.0-1.0) (x.x) and MOVABLEATOMS.

  • NO KEYWORD INFO FOR AMH :( - Mike?
  • AMCHNMAX: The maximum number of angles that will be changed by up to STEP during an AMBER dihedral step. If this is not set or is set to zero, cartesian steps of maximum size STEP are taken instead.
  • AMCHNMIN: The minimum number of angles that will be changed during an AMBER dihedral step.
  • AMCHPMAX: The maximum probability for a single angle to be twisted in an AMBER dihedral step.
  • AMCHPMIN: The minimum probability for a single angle to be twisted in an AMBER dihedral step.
  • ANGSTROM: specifies coordinates in ngstrom for the FRAUSI potential.
  • ARGON: introduces a diatomics-in-molecules calculation for a neutral, cationic or electronically excited argon cluster. See also GROUND, PLUS, TWOPLUS and STAR.
  • ARM arma armb: use the acceptance-ratio method (Bouzida et al., Phys.Rev.A, 45, 8894, 1992) to adjust the step size to achieve the requested acceptance ratio.

A scaling factor is calculated and applied to STEP, rotmax, and/or transmax. The scaling factor is calculated according to where defines the target acceptance ratio and the actual acceptance ratio. Both values arma and armb default to 0.4.

  • ARNO: specifies a diatomics-in-molecules potential for Ar-NO clusters.
  • AVOID dist maxsave: specifies that the geometry should be reseeded if the latest structure gets within a distance dist of the maxsave members of a cyclic list.
  • AXTELL zstar: specifies an additive Axilrod-Teller term for certain diatomics-in-molecules potentials as well as the Pacheco-Ramelho intermolecular potential for C.[2]zstar is the coefficient multiplying this term.
  • BASIN bgmax: specifies a basin-hopping run (as opposed to standard MC on the untransformed surface). bgmax is the convergence threshold on the RMS force in the basin-hopping quenches. If this criterion is too strict then the run time will be greatly increased. If it is too sloppy then the performance of the algorithm is impaired. Different values are needed for different potentials. SLOPPYCONV can be used instead.
  • BFGS: specifies that the full BFGS minimiser should be used. Inefficient compared to LBFGS.
  • BHPT pttmin pttmax exchprob: specifies minimum (pttmin) and maximum (pttmax) temperatures for a parallel tempering basin-hopping run and the probability of attempting replica exchange (exchprob). Should be used together with the MPI keyword. (Only available if the source is compiled with MPI enabled.)
  • BINARY ntypea epsab epsbb sigmaab sigmabb: specifies a binary Lennard-Jones system. ntypea is the number of type A atoms - the rest are assumed to be type B and appear at the end of the list of coordinates. define the units of energy and length, and epsab=, epsbb=, sigmaab=, sigmabb=. The box parameters and cutoff should be specified with the PERIODIC keyword.
  • BINSTRUCTURES SaveNth: requests that the geometry of every SaveNth new structure found during basin-sampling is recorded in binstructures.j, where j is the index of the bin to which a given minimum belongs. If this keyword is present then GMIN switches from plain PTMC to BSPT. Without BINSTRUCTURES the BSPT keyword will perform a standard PTMC run with no quenching.
  • BLJCLUSTER ntypea epsab epsbb sigmaab sigmabb cutoff: specifies a binary Lennard-Jones cluster. The parameters are the same as for BINARY, above.
  • BLN : specifies a BLN off-lattice protein model with bond-length and bond-angle force constants and . An auxiliary file BLNsequence is required.
  • BSMIN: specifies a Bulirsch-Stoer minimisation scheme. Very inefficient compared to LBFGS.
  • BSPT histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nquench nenrper hbins qfrq: requests a basin-sampling run to accumulate the quench probability for local minima as a function of potential energy using a parallel-tempering algorithm.

This keyword also specifies the energy range for the histogram of quench energies, histmin to histmax, the energy range for the histogram of instantaneous configurations, ptemin to ptemax, the temperature range (pttmin and pttmax), the probability of attempting an exchange exchprob, the number of equilibration steps, nequil, the number of parallel tempering MC steps without quenching, ptsteps, the number of parallel tempering MC steps with quenching, nquench, the number of bins for the histogram of instantaneous potential energy, nenrper, the number of bins for the histogram of quench energies, hbins, and the quench frequency, qfrq. Should be used together with the MPI keyword. (This option is only available if the source is compiled with an MPI enabled.)

  • BSPTDUMPFRQ n: n is the interval at which intermediate statistics and BSPTRESTART files are dumped. If n is less than one these files will only be dumped at the end of a complete run. See also BSPTRESTART. Note BSPTRESTART is NOT documented.
  • BSPTDUMPFRQ: restart a previous BSPT or PTMC run. The instantaneous and quench potential energy histograms are read from the last Visits.his and Visits2.his files, and the current state from bsptrestart files (one per node, numbered from zero). A finished run can be continued with more steps by changing the nquench or ptsteps parameters on the BSPT or PTMC line of the data file. Setting the interval for BSPTDUMPFRQ to

minus one will read the last set of dump files.

  • BSWL: obsolete Wang-Landau basin-sampling; do not use.
  • CAPSID rho epsilon radius height: specifies a coarse-grained potential to represent virus capsid pentamers with parameters , , and , respectively. If is omitted the default is 0.5.
  • CENTRE: if present the system will be translated so that the centre-of-mass lies at the origin after every quench.
  • CENTREXY: if present the system will be translated so that the centre-of-mass lies at the centre of the xy plane after every quench. This is useful when using an implicit membrane like IMM1 where you have directionality only in the z-direction, so centreing in x and y should have no delaterious effect.
  • CG: specifies a conjugate-gradient minimisation scheme. Inefficient compared to LBFGS.
  • CHANGEACCEPT naccept: naccept is an integer which sets the interval at which the acceptance ratio is checked and possible adjustments are made to the maximum step size or the temperature. The default is naccept.
  • CHMD chmdfreq: Requests Molecular Dynamics (MD) runs to be performed every chmdfreq steps to generate new geometries. A chmdfreq setting of 20 will execute an MD run every 20 step, while dihedral or cartesian moves are applied otherwise as specified in the data file. A CHARMM parameter file named chmd.par containing all relevant keywords for the CHARMM DYNA module has to be present in the working directory. All CHARMM keywords must be uppercase and given in the first line. A typical example is:

VERL NSTEP 500 TIMESTEP 0.002 TWINDH 10.0 IEQFRQ 200 ICHECW 1 IASORS 0 IASVEL 1 FIRS 500 FINA 500

Please consult the CHARMM manual for further details on the DYNA module. Currently, the length of the input string given in chmd.par is limited to 500 characters.

  • CHARMMENERGIES: prints the components of the total CHARMM energy after each step.
  • CHARMMTYPE topfile paramfile: topfile and paramfile are the common CHARMM top and param files, e.g. 'toph19\_eef1\_perm.inp' and 'param19\_eef1\_perm.inp'.
  • CHFREQ nfreq: used with CHARMM keyword to specify that every nfreq basin-hopping steps dihedrals are twisted. Default is nfreq=1.
  • CHNMAX: used with CHARMM keyword to specify the maximum allowed number of angles to be twisted. Specifies a dihedral angle step taking scheme.
  • CHNMIN: used with CHARMM keyword to specify the minimum allowed number of angles to be twisted.
  • CHPMAX: used with CHARMM keyword to specify the maximum allowed probability for twisting an angle.
  • CHPMIN: used with CHARMM keyword to specify the minimum allowed probability for twisting an angle.
  • CHRIGIDROT prot rotmax nrot: used with CHARMM keyword to support rigid body rotation every nrot basin-hopping steps with maximum allowed probability prot and maximum allowed rotation angle rotmax (in degrees). The keyword CHRIGIDROT requires a file segments.tomove, which specifies the segments for rigid rotation. The segments are numbered and each line contains only one number.
  • CHRIGIDTRANS ptrans transmax ntrans: used with CHARMM keyword to support rigid body translation every ntrans basin-hopping steps with maximum allowed probability ptrans and maximum allowed translation transmax (in ). The keyword CHRIGIDTRANS requires a file segments.tomove, which specifies the segments for rigid translation. The segments are numbered and each line contains only one number. CHRIGIDROT and CHRIGIDTRANS use the same segments.tomove.
  • CISTRANS: disables all checks for cis or deformed amide/peptide bonds.
  • COLDFUSION thresh: if the energy falls below threshold thresh then cold fusion is assumed to have occurred and geometry optimisation stops. The default value is .
  • COMPRESS comp: add a harmonic compression potential with force constant comp using the centre-of-mass distance for each atom.
  • COMMENT: the rest of the line is ignored.
  • COOPMOVE n cut: specifies cooperative moves in the step-taking routine. An atom is selected at random, and the n nearest neighbours (default 5) that lie within a cutoff distance of cut (default 1.0) are moved by the same amount.
  • CPMD sys: specifies that the CPMD program should be called for energies and gradients. Not tested!
  • CUTOFF cutoff: sets a cutoff beyond which the potential is truncated. This only has an effect for tight-binding silicon at present.
  • DBRENT specifies minimisation using Brent's method with first derivatives in the conjugate-gradient procedure. Inefficient compared to LBFGS.
  • DEBUG: sets various debug printing options including the dumping of initial geometries and energies (to dump.X.xyz) if DUMP is also set.
  • DECAY x: magnitude of random move decays according to parameter x with distance from a randomly chosen atom.
  • DF1: specifies a binary 2D potential. The first atoms have unit radius and the rest have radius 1.4, with a cutoff for each pair type at the average radius. The keyword 2D must also be specified, along with a PERIODIC line to specify two box-lengths. Initial work uses box lengths of 3.31437171 for a number density of 0.9.
  • DFTB: specifies a DFT-based tight-binding potential; the multiplicity is specified by keyword MULTIPLICITY.
  • DGUESS dguess: initial guess for diagonal elements of the inverse Hessian, used whenever the LBFGS optimiser is reset. The default is dguess=0.1.
  • DIELEC dparam: specifies dielectric constant for AMBER - old implementation.
  • DIPOLES: causes the first order induction energy to be included in a diatomics-in-molecules calculation for Ne or Ar. By default this term is neglected, although it may be significant.
  • DUMP: if present will cause the energy and quench geometry for every step to be dumped into dump.X.xyz where X is an integer. The geometries are saved in XMakemol xyz format. If CHARMM is also specified, dump.pdb and dump.crd are produced containing each quench geometry in PDB and CHARMM CRD format.
  • DUMPINT int: changes the default interval for dumping a restart GMIN.dump file from 1000 basin-hopping steps to int.
  • DUMPQU: when using AMBER9, dumps each quench geometry in rst format to quenchX.rst and pdb format to quenchX.pdb. Dumping does not occur if a chirality check fails.
  • DZUGUTOV dzp1 dzp2 dzp3 dzp4 dzp5 dzp6 dzp7: Dzugutov potential in a general form. The parameters are , , , , , and .
  • EAMAL: specifies an embedded atom model for aluminium.
  • EAMLJ A0 beta Z0: specifies the EAMLJ potential (Baskes, Phys.Rev.Lett., 27, 2592, 1999) with parameters A0, beta and Z0.
  • EDIFF econv: quench minima are only considered to be different if their energies differ by at least . This option mainly affects the lowest energy saved geometries. If the current quench energy is within of a saved energy, but lies lower, then the saved energy and geometry are replaced. The default is but different values are appropriate for different potentials.
  • EQUILIBRATION equil DumpEveryNthQuench: equil is the number of MC steps preceding the accumulation of the density of states histogram in a Wang-Landau basin-sampling run. The default is 0. DumpEveryNthQuench specifies how often the statistics are recorded into the output files.
  • FAKEWATER: specifies a distance-dependent dielectric in AMBER - old implementation.
  • FAL: specifies the Farkas potential for aluminium.
  • FIXBOTH: both the temperature and maximum step size are fixed regardless of the calculated acceptance ratio.
  • FIXCOM: fix centre of mass rather than centre of coordinates.
  • FIXSTEP: the maximum step size is fixed and the temperature is varied to try and achieve the requested acceptance ratio.
  • FNI: specifies the Farkas potential for nickel.
  • FRAUSI: specifies a particular tight-binding potential for silicon. See also keyword ANGSTROM.
  • FREEZE n1 n2 : freeze the coordinates of atoms n1, n2,}. Note that all the FREEZE keywords do NOT function when using AMBERMDSTEPS. Use multiple FREEZE statements for larger numbers of atoms.
  • FREEZERES n1 n2 : freeze the coordinates of all atoms in residues n1, n2,.
  • FREEZEALL n1 n2 : freeze the coordinates of all atoms.
  • FS gatom: specifies a Finnis-Sinclair potential using parameters from Finnis and Sinclair, Phil.Mag.A, 50, 45 (1984) and corresponding erratum Phil.Mag.A, 53, 161 (1986). gatom=1 for V, gatom=2 for Nb, gatom=3 for Ta, gatom=4 for Cr, gatom=5 for Mo, gatom=6 for W, gatom=7 for Fe (original parameters), gatom=8 for Fe (modified parameters in erratum). Subtoutine FS was coded by James Elliott in April 2009.

G-L

  • GROUND: when combined with keywords NEON or ARGON uses an accurate (Aziz) potential to model the ground state neutral cluster.
  • GUIDE guidecut: specifies the RMS force below which the real potential is used rather than a guiding potential. The systems affected are CPMD and WELCH, which are guided by AMBER and TOSI, respectively, and also PACHECO, where the Axilrod-Teller contribution is only included when the RMS force falls below guidecut. Default guidecut=0.0001. New guided potentials are ZETT1 and ZETT2 (guided by Morse with ) and NATB (guided by GUPTAT). Parameters for the guiding potential must also be specified in data.
  • GUPTA gatom: specifies a Gupta potential using parameters from Cleri and Rosato, Phys.Rev.B, 48, 22 (1993). gatom=1 for Ni, gatom=2} for Cu, gatom=3 for Rh, gatom=4 for Pd, gatom=5 for Ag, gatom=6 for Ir, gatom=7 for Pt, gatom=8 for Au, gatom=9 for Al, gatom=10 for Pb, gatom=11 for Ti type 1, gatom=12 for Ti type 2, gatom=13 for Zr type 1, gatom=14 for Zr type 2, gatom=15} for Co, gatom=16 for Cd type 1, gatom=17 for Cd type 2, gatom=18 for Zn, gatom=19 for Mg, gatom=20 for V, gatom=21 for Na, gatom=22 for Sr (Wang and Blaisten-Barojas, J.Chem.Phys., 115, 3640 (2001)), gatom=? for Au as used by Garzon et al. The Gupta subroutine was recoded more efficiently by James Elliott in April 2009.
  • INTMIN: used with CHARMM keyword to specify minimisation in internal coordinates. This generally appears to be slower than using Cartesian coordinates.
  • JC: Specifies Murrell's two-plus three-body potential.[3][4][5][6][7] A file {\tt JMparams} must exist in the current directory containing the parameters , , , , , , and . An optional cutoff parameter can also be provided at the end of the JMparams file. Subroutines used: jmec, jm2c, jm3c.
  • JUMPMOVE np1 np2 int: specify J-walking type attempts between parallel runs np2 and np1 at intervals of int steps.
  • LB2: specifies the potential [8][9][10], where and are set to unity.
  • LIGMOVE ligrotscale ligcartstep: used with AMBER9 and MOVABLEATOMS. Specifies ligand only rotation and cartesian perturbations. The ligand is defined by atom index in the file movableatoms. Setting ligrotscale less than 1.0 limits the ammount of rotation possible - this may be required to prevent cold fusion with non-spherical ligands. ligcartstep defines the maximum size (in angstroms) of the random cartesian perturbations applied to the ligand. Both rotation and cartesian ligand moves are applied AFTER any MD if AMBERMDMOVES is on to prevent the MD exploding.
  • LJCOUL nc f : specifies a cluster of Lennard-Jones particles in which the first nc particles carry identical reduced charges in addition to the Lennard-Jones interaction. The parameter f specifies what fraction of the Monte Carlo steps should be swaps between the positions of a charged and a neutral particle, rather than a conventional step. is the temperature to be used in the acceptance criterion for swap moves, overriding that specified using the TEMPERATURE keyword. Generally, a lower temperature is more effective at finding the lowest-energy permutation of charges. The default value of is zero. The reduced charge is related to the actual charge by , where and are the Lennard-Jones well depth and length parameter respectively. This way, the reduced energy of two charges is , where is the reduced distance between the charges.
  • LOCALSAMPLE abthresh acthresh: Keyword currently under construction! For three groups of atoms defined in movableatoms (A,B,C), a step is quenched when the centre of coordinates of A->B is less than abthresh AND A->C is less than acthresh. If this condition is broken AFTER the quench, it is automatically rejected.

M-R

  • MAXBFGS max: max is the largest permitted LBFGS step.
  • MAXERISE maxez: specifies the largest rise in energy permitted during an LBFGS minimisation, default . Useful for potentials with imprecise derivatives.
  • MAXIT maxit maxit2: maxit and maxit2 are integers specifying the maximum number of iterations allowed in the conjugate gradient quenches. maxit applies to the 'sloppy' quenches of the basin-hopping run and maxit2 to the final quenches that are used to produce the output in file lowest.
  • MGGLUE: specifies a glue potential for magnesium.
  • MPI: specifies an MPI parallel job. (only available if the source is compiled with MPI enabled).
  • MSORIG: specifies a particular tight-binding potential for silicon.
  • MSTRANS: specifies an alternative tight-binding potential for silicon.
  • MULTIPLICITY xmul: specifies the multiplicity of the electronic state in DFTB calculations.
  • NATB: specifies the sodium tight-binding potential of Calvo and Spiegelmann. This potential can be guided by also specifying GUPTA 21 in the data file.
  • NEON: introduces a diatomics-in-molecules calculation for a neutral, cationic or electronically excited neon cluster. See also GROUND, PLUS, TWOPLUS and STAR.
  • NEWJUMP prob: for (serial) 'parallel' runs specifies a jump probability between runs (parallel tempering) of prob. See the BHPT keyword for a better alternative.
  • NEWRESTART nrelax nhs MD newrestemp: reseed runs if the energy does not decrease within nrelax steps. nhs is the number of hard sphere moves used to produce the new starting configuration. If nhs=0 (the default) then the geometry is changed by reseeding. If MD is present, then a short AMBER or CHARMM MD run is performed at temperature newrestemp to generate the new configuration.
  • NMAX nmax: nmax is an integer that specifies the maximum number of dihedral angles to be twisted in AMBER - old implementation.
  • NMIN nmax: nmin is an integer that specifies the minimum number of dihedral angles to be twisted in AMBER - old implementation.
  • NOCHIRALCHECKS: disables checks for inversion of CA atoms and chiral side-chains for ILE and THR.
  • NOCISTRANS minomega: set on by default with a threshold minomega of 150 degrees. If an amide bond is deformed to a angle below the specified threshold, the structure is discarded. i.e. with minomega. minomega defaults to 150 degrees. For proline every is allowed. To enable cis-trans isomerisation, the CISTRANS keyword should be used.
  • NOCISTRANSDNA minomega: should be specified when working with DNA in AMBER to ensure the correct bonds are checked. As above, the deformation threshold minomega can be set. It is defaulted to 150 degrees.
  • NOCISTRANSRNA minomega: should be specified when working with RNA in AMBER to ensure the correct bonds are checked. As above, the deformation threshold minomega can be set. It is defaulted to 150 degrees.
  • NOFREEZE: don't freeze the core atoms when doing the initial geometry optimisations in a run where SEED is specified.
  • NOPHIPSI: used with the CHARMM keyword to specify twisting of sidechain dihedrals only.
  • NORESET: by default the configuration point is set to that of the quench minimum in the Markov chain during a basin-hopping simulation. This keyword turns off the resetting so that the geometry varies continuously.
  • NOTE: the rest of the line is ignored.
  • ODIHE: order parameter - requires documentation.
  • OEINT: interaction energy between 2 peptides will be used as an order parameter - requires further documentation.
  • ORGYR: radius of gyration will be calculated as an order parameter - requires further documentation.
  • OSASA: order parameter - requires documentation.
  • P46: specifies a 46-bead three-colour model polypeptide. See also the BLN keyword, which implements this potential in a more general way and uses unit bond lengths.
  • PACHECO: specifies the intermolecular Pacheco-Ramelho potential for C. The Axilrod-Teller contribution, specified with the AXTELL keyword, is included when the RMS force falls below the value entered with guidecut.
  • PAH: specifies a polycyclic aromatic hydrocarbon potential.
  • PARALLEL npar: npar is the number of parallel runs within GMIN.
  • PBGLUE: specifies a glue potential for lead.
  • PERIODIC boxlx boxly boxlz: specifies periodic boundary conditions for potentials which understand such a directive (such as tight-binding silicon). The three double precision variables are the box lengths. If only one box length is given the others are set to the same value to give a cube.
  • PERMDIST: minimise distances between the coordinates in files coords and the fixed coordinates in file finish with respect to permutational isomerisation. Requires the auxiliary file perm.allow to specify permutable atoms, otherwise all atoms are assumed to be permutable. The absence of a perm.allow file is considered a mistake for CHARMM runs. The first line of the perm.allow file must contain an integer that specifies the number of groups of interchangeable atoms. The groups then follow, each one introduced by a line with two integers that specify the number of permutable atoms and the number of other pairs of atoms that must swap if the first pair are permuted. The following line contains the numbers of the permutable atoms and then the numbers of the swap pairs, in order. For CHARMM runs the perm.allow file can be generated automatically from the input.crd file using the python script perm.py written by Dr Mey Khalili. Note that it is then essential to use CHARMM topology files that include symmetrised definitions of all the relevant amino acids if energies are actually calculated.
  • PLUS: when combined with keywords NEON or ARGON uses a diatomics-in-molecules potential for the singly charged cation.
  • PMAX max: max is the maximum probability of dihedral angle twisting for AMBER - old implementation?.
  • PMIN min: min is the minimum probability of dihedral angle twisting for AMBER - old implementation?.
  • POWER ipow: ipow is the initial power for shifts in the old line minimisation routine for conjugate gradient. LBFGS minimisation should be used instead.
  • PROJI: turns on projection operator to enforce point group symmetry in mylbfgs.f. The geometry is projected after every proposed step.
  • PROJIH: turns on projection operator to enforce point group symmetry in mylbfgs.f. The geometry is projected after every proposed step.
  • PTMC histmin histmax ptemin ptemax pttmin pttmax exchprob nequil ptsteps nenrper hbins: requests a standard parallel tempering MC run.

This keyword also specifies the energy range for the histogram of quench energies, histmin to histmax, the energy range for the histogram of instantaneous configurations, ptemin to ptemax, the temperature range (pttmin} and pttmax), the probability of attempting an exchange exchprob, the number of equilibration steps, nequil, the number of parallel tempering MC steps without quenching, ptsteps, the number of bins for the histogram of instantaneous potential energy, nenrper, and the number of bins for the histogram of quench energies, hbins. Should be used together with the MPI keyword. (This option is only available if the source is compiled with an MPI enabled.)

  • PULL a1 a2 f: apply a static force to the potential, equivalent to adding the term . Here and are the coordinates for atoms and , and specifies the force. This potential is designed to simulate a pulling experiment with static force where a molecule is pulled along the axis from atoms and .
  • QCUTOFF qcut: qcut is a distance cut-off for Coulomb interactions in AMBER.
  • QMAX cgmax: cgmax is the tolerance for the RMS force in the final set of quenches that are used to produce the output for file lowest. The default is cgmax, but the appropriate value depends upon the system in question. TIGHTCONV can be used instead.
  • QUAD: requires documentation :(.
  • QUCENTRE: sets the centre of coordinates to the origin (0,0,0) before each MC step is taken (so after each quench), but not during the minimisation itself unlike CENTRE.
  • RADIUS radius: sets the radius of the container that prevents particles evaporating during quenches.

If unset the program calculates an appropriate value based upon the volume per particle for close-packed material and the known pair equilibrium distance for the given potential. The formula employed is:

where is the number of atoms and is the pair equilibrium separation. [14] The '1' in this formula is to allow some extra space for more open structures.

  • RANDOMSEED: specifies that the random number generator should be seeded with system time after each quench, allowing simple parallel use. Currently functional only for the CHARMM and AMBER potentials.
  • RANSEED i: integer seed for the random number generator.
  • RCUTOFF rcut: rcut is a distance cut-off in AMBER - old implementation.
  • RESIZE resize: all the coordinates are multiplied by resize after they have been read in, before any other operations are performed. This command is useful for scaling results obtained with one potential for a system with a different pair equilibrium distance.
  • RESTART nrelax nhs: reseed runs if a step in not accepted within twice nrelax steps. nhs is the number of hard sphere moves used to produce the new starting configuration. If nhs=0 (the default) then the geometry is changed by reseeding.
  • RESTORE dumpfile intEdumpfile: restore a previous GMIN run from a dumpfile. The number of basin-hopping steps performed will be the difference between the number requested for the run that produced the dumpfile, minus the number that were completed at the point the dumpfile was created. This option is not available before version 2.3. If you are using the A9INTE keyword, you can specify the interaction enthalpy dump file to restore from as a second arguement.
  • RGCL2: specifies a DIM rare gas-Cl potential.
  • RINGROTSCALE factor: when applying cartesian moves with CHARMM, amino acid rings are moved as rigid units. Setting factor (default 0.0) between 0.0 and 1.0 will apply a random rotation to these rings during step taking. The suggested value is 0.1 to prevent the regular formation of high energy structures.
  • RKMIN: specifies a Runga-Kutta minimisation scheme. Very inefficient.
  • RMS rmslimit rmstol rmssave lca best: used with CHARMM keyword to specify that the RMSD compared to a reference geometry is calculated.

The reference geometry must be given in xyz-format in an additional file compare. rmssave is an integer that specifies the number of lowest energy geometries and RMSD rmslimit to save. Geometries are only saved if their RMSD's are more than rmstol different. The flag lca controls whether the all-atom RMSD (lca=0) or the -RMSD (lca=1) is calculated. The flag best determines which structure is compared to the reference after each quench. best=0 implies the current quench minimum and best=1 implies the current best (lowest energy) minimum. If TRACKDATA is also specified, the RMSD calculated after each quench is produced in the file rmsd in gnuplot readable format.

S-Z

  • SAVE nsave: nsave is an integer that specifies the number of lowest energy geometries to save and summarise in the file lowest. Arrays are now dynamically allocated, so any positive integer can be specified.
  • SAVEINTE nsaveinte}: nsaveinte is an integer that specifies the number of lowest interaction enthalpy geometries to save and summarise in the file intelowest. See A9INTE.
  • SETCENTRE x y z: Sets the centre of mass/coordinates (before the initial quench) to (x,y,z). For example, SETCENTRE 0.0 0.0 0.0 would translate the centre of mass to the origin.
  • SC nn mm sig sceps scc: specifies a Sutton-Chen potential [15] with parameters nn, mm, =sig, sceps and scc.
  • SEED nsstop: if the SEED keyword appears then the program looks for a file seed containing coordinates, which are used to 'seed' the new run. The number of coordinates given in this file should be no more than one less than the number given in coords. The specified coordinates are frozen from the first step until step nsstop.
  • SHIFTCUT: specifies a shifted-truncated potential for bulk binary Lennard-Jones.
  • SIDESTEP smax: specifies the maximum step in Cartesian coordinates for side-chains

in AMBER - old implementation.

  • SLOPPYCONV bgmax: specifies a basin-hopping run (as opposed to standard MC on the untransformed surface). bgmax is the convergence criterion for the RMS force in the basin-hopping quenches. If this criterion is too strict then the run time will be greatly increased. If it is too sloppy then the performance of the algorithm is impaired. Different values are needed for different potentials. BASIN can be used instead.
  • SORT: for pairwise potentials the atoms can be sorted from most to least strongly bound. The SORT keyword enables this sorting for the coordinates printed in file lowest. This can be useful for seeding subsequent runs by removing the most weakly bound atoms. This sort is not set by default and is meaningless if the pair energies are not computed.
  • STAR: specifies an excited state calculation for Ar or Ne for a diatomics-in-molecules potential when used with NEON or ARGON.
  • STEP step astep ostep block: specifies the maximum step sizes.

step is for the maximum change of any Cartesian coordinate and astep specifies a tolerance on the binding energy of individual atoms (if available, i.e. for Morse and LJ) below which an angular step is taken for that atom. ostep is the maximum displacement of an axis-angle coordinate for a rigid body system and block (an integer) is the block size for which separate translational and orientational displacements will be made for rigid bodies. Omitting block or using a value of zero results in translational and orientational steps being taken simultaneously for rigid bodies. The default values for step, astep and ostep are all 0.3 and the default value of nblock is zero.

  • STEEREDMIN smink sminkinc smindiststart smindistfinish sminatoma sminatomb: specified steered minimisation should be performed (must be used with AMBER9).

For a protein/ligand system, this adds a translation to the MC move. The vector between the centre of coordinates of groups A and B (as defined in the file movableatoms) is calculated and set to smindiststart. During the following minimisation, a restoring force is applied to the ligand. The harmonic force constant is initially zero, and rises by sminkinc every LBFGS step up to a maximum of smink. The force is applied until the A->B distance is less than smindistfinish.

  • STEPS mcsteps tfac: determines the length of the basin-hopping run through the integer mcsteps and the annealing protocol through the real variable tfac. The temperature is multiplied by tfac after every step in each run.
  • STICKY nrbsites, sigma: specifies a 'sticky patch' potential with nrbsites sites in the rigid body reference and a value of sigma for the parameter.
  • STOCK mu lambda: specifies a Stockmeyer potential with parameters and , respectively.
  • STRAND: specifies a system of strands coded using the rigid body formalism.
  • SW: specifies the Stillinger-Weber Si potential.
  • SYMMETRISE int tol1 tol2 tol3 tol4 tol5 qmax mdiff d: specifies that the symmetrisation routine should be called every int steps.

The five tol parameters are tolerances for various parts of the routine: tol1 is used in ptgrp.f in defining orbits; tol2 is the distance tolerance used in ptgrp.f to define point group symmetry operations; tol3 is the maximum relative difference in principal moments of inertia used to diagnose point groups with degenerate irreducible representations in ptgrp.f; tol4 is the distance cutoff used to determine if a symmetry element has been lost in symmetry.f. Since we are dealing with approximate symmetries, this parameter may be larger than tol2. It is compared to the largest atomic displacement divided by the corresponding radius for the closest permutation. tol4 is also used to test whether atoms lie on a given symmetry element, and in testing whether orbits generated from 'floaters' are actually contained in the core. tol5 is generally to check for atom clashes in symmetry.f, including analysis of missing sites in orbits, as well as overlap between orbits generated from 'floaters' and previous core or new orbit sites. qmax is the maximum number of quenches allowed for each call to symmetry.f. mdiff is used to test whether a generated symmetry operation is new. If any of the nine components of the corresponding matrix differs by more than mdiff from an existing matrix then the operations are considered to be different. d is the exponential factor used in constructing a centre of mass that is biased towards core atoms. The contribution of each atom is weighted by , where is the centre of mass distance of atom on the previous cycle.

  • TABOO nlist: specifies a taboo list of the nlist lowest minima should be maintained.
  • TARGET target1 target2 : specifies any number of target energies. The current run stops in an orderly fashion if the current quench energy is within econv of any target (see EDIFF).
  • TEMPERATURE temp: defines the temperature, temp, at which the

MC runs are conducted. Different values can be specified for serial 'parallel' runs if PARALLEL is set. For true parallel basin-hopping use the BHPT keyword and omit TEMPERATURE.

  • TETHER hdistconstraint hwindows ExtrapolationPercent lnHarmFreq: requests a calculation of the vibrational density of states for a given minimum. hdistconstraint is the minimised average radius of the basin of attraction to which the minimum belongs, hwindows is the number of potential energy windows into which a WL simulation is split. ExtrapolationPercent is the percentage of the whole potential energy spectrum, for which the density of states is estimated from the harmonic approximation and not sampled. lnHarmFreq the log product of positive Hessian eigenvalues.
  • THOMSON q: specify the Thomson problem for unit charges on a sphere. If q is present it is taken to be the charge on one particle, which can therefore be different from all the other unit charges and is read as a real number.
  • TIGHTCONV cgmax: cgmax is the tolerance for the RMS force in the final set of quenches that are used to produce the output for file lowest. The default is cgmax, but the appropriate values depend upon the system in question. QMAX can be used instead.
  • TIP n: specifies a TIPnP intermolecular potential for rigid body water molecules. .
  • TOLBRENT tolb: parameter for DBRENT minimisation. Inefficient compared to LBFGS.
  • TOMEGA: used with the CHARMM keyword to specify that peptide bonds will be twisted along with all other dihedrals.
  • TOSI app amm apm rho: specifies the Tosi-Fumi potential [16] with parameters , , and .
  • TRACKDATA: produces energy.dat and markov.dat containing the quench number and associated energy and markov energy in two columns and best.dat, containing the current quench number and the current lowest total energy. If RMS is also specified, a file called rmsd.dat is produced containing the RMSD from a reference structure. See RMS for more information. This allows plotting with gnuplot to monitor convergence of multiple runs. If A9INTE is also specified, two additional output files are produced, intE.dat containing the quench number and associated interaction enthalpy, and bestintE.dat containing the quench number and current lowest interaction enthalpy. This keyword does not yet function for MPI runs.
  • TSALLIS q: specifies that steps are accepted/rejected using Tsallis statistics with the given value of q, rather than the usual Boltzmann condition.
  • TWOPLUS: when combined with keywords NEON or ARGON uses a diatomics-in-molecules potential for the doubly charged cation.
  • UACHIRAL: MUST be included when using ff03ua, the AMBER united atom forcefield unless you have disabled the checks for inverted chiral carbons with NOCHIRALCHECKS. UACHIRAL ensures the correct impropers are used to define sidechain chirality when HB hydrogen is missing.
  • UNFREEZE n1 n2 : unfreeze the coordinates of atoms n1, n2,. Only functions in conjunction with FREEZEALL.
  • UNFREEZERES n1 n2 : unfreeze the coordinates of all atoms in residues n1, n2,. Only functions in conjunction with FREEZEALL.
  • UPDATES nup: nup is the number of previous steps saved in the LBFGS routine,

default 4.

  • VGW ljsigma ljepsilon taumaxsg taumaxfg: Specifies use of VGW quantum quenching in place of classical minimization routines such as LBFGS.

ljsigma and ljepsilon are the corresponding Lennard-Jones parameters that must be specified, and taumaxsg and taumaxfg are the maximum value of 'imaginary' time (inverse tempertaure) for the propagation. The former pertains to the faster 'single-particle' SP-VGW used for quenching during the MC runs, and the latter for the more accurate 'fully-coupled' VGW used for the final quenching (analogous to the tight convergence of the LBFGS). A of at least 2.5 is recommended for the SP-VGW and 5.0 for the FC-VGW. A file vgwdata containing the masses (in a.m.u.) of all particles, in order of the location of their xyz coordinates in coords must be present (e.g. for a 38 atom Ne cluster, vgwdata will have 38 lines of '20'). Different masses are permitted, though the current version allows for only one set of LJ parameters.

  • VGWCPS on magnitude}: Specifies use of contraining potential for SP-VGW (sloppy convergence), as clusters expand during quantum quenching with decreasing mass. 1 or 0 for on corresponds to on/off, and magnitude should range from 1 to 1000, with 1 having minimal effect, 1000 being highly constrained. Default value is on, with magnitude 1.
  • VGWCPF on magnitude: Same as VGWCPS but for FC-VGW, used for the final, full quenching (tight convergence).
  • VGWTOL magnitude: Absolute tolerance parameter for differential equation solver used for VGW quenching. Default value is 0.0001. For highly quantum or 'stiff' systems this may need to be increased, while it may be decreased for 'softer' or less quantum systems to enhance speed.
  • VISITPROP: if specified the Wang-Landau convergence is governed by proportionality of visits to the current value of the modification factor, and not the histogram flatness criterion [17].
  • WELCH : specifies a Welch binary salt potential with the parameters indicated.
  • ZETT1 and ZETT2: specify the Zetterling potentials.

Some recognised systems

Bibliography

  1. Bogdan, T. V., Wales, D. J., Calvo, F. - Equilibrium thermodynamics from basin-sampling (13 pages).
    J. Chem. Phys. 124:044102,2006
    Bibtex
    Author : Bogdan, T. V., Wales, D. J., Calvo, F.
    Title : Equilibrium thermodynamics from basin-sampling (13 pages).
    In : J. Chem. Phys. -
    Address :
    Date : 2006
  2. Pacheco, J. M., Ramalho, J. P. P. - First-principles determination of the dispersion interaction between fullerenes and their intermolecular potential
    \prl 79:3873,1997
    Bibtex
    Author : Pacheco, J. M., Ramalho, J. P. P.
    Title : First-principles determination of the dispersion interaction between fullerenes and their intermolecular potential
    In : \prl -
    Address :
    Date : 1997
  3. Murrell, J. N., Mottram, R. E. - potential energy functions for atomic solids
    \molphys 69:571-585,1990
    Bibtex
    Author : Murrell, J. N., Mottram, R. E.
    Title : potential energy functions for atomic solids
    In : \molphys -
    Address :
    Date : 1990
  4. Murrell, J. N., Rodriguez-Ruiz, J. A. - potential energy functions for atomic solids 2. potential functions for diamond-like structures
    \molphys 71:823-834,1990
    Bibtex
    Author : Murrell, J. N., Rodriguez-Ruiz, J. A.
    Title : potential energy functions for atomic solids 2. potential functions for diamond-like structures
    In : \molphys -
    Address :
    Date : 1990
  5. Alderzi, A. R., Johnston, R. L., Murrell, J. N., Rodrigez-Ruiz, J. A. - potential energy functions for atomic solids. 3. fitting phonon spectra and elastic constants of diamond structures
    \molphys 73:265-282,1991
    Bibtex
    Author : Alderzi, A. R., Johnston, R. L., Murrell, J. N., Rodrigez-Ruiz, J. A.
    Title : potential energy functions for atomic solids. 3. fitting phonon spectra and elastic constants of diamond structures
    In : \molphys -
    Address :
    Date : 1991
  6. Eggen, B. E., Johnston, R. L., Li, S., Murrell, J. N. -
    \molphys 76:1405,1992
    Bibtex
    Author : Eggen, B. E., Johnston, R. L., Li, S., Murrell, J. N.
    Title :
    In : \molphys -
    Address :
    Date : 1992
  7. Feng, J.-Y., Johnston, R. L., Murrell, J. N. -
    \molphys 78:1405,1993
    Bibtex
    Author : Feng, J.-Y., Johnston, R. L., Murrell, J. N.
    Title :
    In : \molphys -
    Address :
    Date : 1993
  8. Lynden-Bell, D., Lynden-Bell, R. M. -
    Proc. Roy. Soc. Lond. A 455:475-489,1999
    Bibtex
    Author : Lynden-Bell, D., Lynden-Bell, R. M.
    Title :
    In : Proc. Roy. Soc. Lond. A -
    Address :
    Date : 1999
  9. Lynden-Bell, D., Lynden-Bell, R. M. -
    Proc. Roy. Soc. Lond. A 455:3261-3284,1999
    Bibtex
    Author : Lynden-Bell, D., Lynden-Bell, R. M.
    Title :
    In : Proc. Roy. Soc. Lond. A -
    Address :
    Date : 1999
  10. Lynden-Bell, D., Lynden-Bell, R. M. -
    J. Stat. Phys. 117:199-209,2004
    Bibtex
    Author : Lynden-Bell, D., Lynden-Bell, R. M.
    Title :
    In : J. Stat. Phys. -
    Address :
    Date : 2004
  11. Braier, P. A., Berry, R. S., Wales, D. J. - How the range of pair interactions governs features of multidimensional potentials
    \jcp 93:8745,1990
    Bibtex
    Author : Braier, P. A., Berry, R. S., Wales, D. J.
    Title : How the range of pair interactions governs features of multidimensional potentials
    In : \jcp -
    Address :
    Date : 1990
  12. Doye, J. P. K., Wales, D. J., Berry, R. S. - THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURES OF CLUSTERS
    \jcp 103:4234-4249,1995
    Bibtex
    Author : Doye, J. P. K., Wales, D. J., Berry, R. S.
    Title : THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURES OF CLUSTERS
    In : \jcp -
    Address :
    Date : 1995
  13. Doye, J. P. K., Wales, D. J. - THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURE AND STABILITY OF SIMPLE LIQUIDS - FROM CLUSTERS TO BULK, FROM SODIUM TO C-60
    J. Phys. B 29:4859-4894,1996
    Bibtex
    Author : Doye, J. P. K., Wales, D. J.
    Title : THE EFFECT OF THE RANGE OF THE POTENTIAL ON THE STRUCTURE AND STABILITY OF SIMPLE LIQUIDS - FROM CLUSTERS TO BULK, FROM SODIUM TO C-60
    In : J. Phys. B -
    Address :
    Date : 1996
  14. Kittel, C. - Introduction to Solid State Physics
    third, Wiley, New York,1976
    Bibtex
    Author : Kittel, C.
    Title : Introduction to Solid State Physics
    In : -
    Address : New York
    Date : 1976
  15. Sutton, A. P., Chen, J. - long-range Finnis-Sinclair potentials
    Phil. Mag. Lett. 61:139,1990
    Bibtex
    Author : Sutton, A. P., Chen, J.
    Title : long-range Finnis-Sinclair potentials
    In : Phil. Mag. Lett. -
    Address :
    Date : 1990
  16. Tosi, M. P., Fumi, F. G. -
    J. Phys. Chem. Solids 25:45,1964
    Bibtex
    Author : Tosi, M. P., Fumi, F. G.
    Title :
    In : J. Phys. Chem. Solids -
    Address :
    Date : 1964
  17. C. Zhou, R.N. Bhatt - Understanding and Improving the Wang-Landau Algorithm
    \pre 72:025701,2005
    Bibtex
    Author : C. Zhou, R.N. Bhatt
    Title : Understanding and Improving the Wang-Landau Algorithm
    In : \pre -
    Address :
    Date : 2005