[ autoSavePeriod | defSymGroup | i_out | i_2out | maxColorPotential | maxFileSizeMb | maxMemory | minTetherWindow | mnSolutions | mncalls | mncallsMC | mnconf | mnhighEnergy | mnreject | mnvisits | nLocalDeformVar | nSsearchStep | nProc | randomSeed | segMinLength | sequenceBlock | sequenceLine | surfaceAccuracy | windowSize ]
Here is the alphabetically sorted dump of integer parameters defined in the ICM-shell. These parameters are used by various commands and functions and can be changed interactively, e.g.mncallsMC= 10000
montecarlo
ICM-shell integer variables are the following.
montecarlo or ssearch
procedures which may run for days, the current stack of
conformations which accumulates the best energy representatives of different
conformational areas is saved periodically to allow access to intermediate
results of the simulations. The above parameter defines the number of
stack changes after which it is saved to a file.
Set autoSavePeriod to 1 if you want
to be conservative.
If you set
autoSavePeriod to 0 , the stack will not be saved at all.
Default (10).
Symgroup function.
Default (0) means that the group is not defined.
Examples:
defSymGroup = 19 # direct assignment. You know group 19, don't you?
defSymGroup = Symgroup("P212121") # Oh, you do not! ..
defSymGroup = Symgroup("P61 2 2") # This one you do not remember for sure
-
Rmsdsaves the number of aligned equivalent points; -
Srmsdsaves the number of aligned equivalent points; -
convertsaves the number of heavy atoms missing from the pdb-template (e.g. atoms of the flexible lys side-chain are not given in the pdb-file). -
superimposesaves number of aligned equivalent points; - set tether saves the number of tethers imposed;
- set drestraint saves the number of distance restraints imposed;
- set vrestraint saves the number of variable restraints imposed;
- make disulfide bond saves the number of imposed disulfide bonds;
-
minimizesaves the number of function evaluations; -
montecarlosaves the total number of function evaluations during minimization; - show area skin saves the total number of triangles in the Connolly construction.
Default (0).
the second variable for additional integer output. (see also i_out and I_out )
If the maxColorPotential is set to 0. the color grob potential
command will perform automated scaling to the absolute maximal value of the
potential.
See also: color grob potential [heavy], dsRebel , Potential , make grob potential , TOOLS.rebelPatchSize ,
Example:
build string "se glu arg" # dipeptide
maxColorPotential = 3.
dsRebel a_ yes
maxColorPotential = 6.
TOOLS.rebelPatchSize = 0.5 # do not do it for very large objects
dsRebel a_ yes
maximal file size in Megabytes that can be read in ICM. Increase this limit if needed. Example:
maxFileSizeMb = 10000
read table mol "giant.sdf"
See also: write index
Recommendation: divide your available RAM by a factor from 2 to 4.
Current memory resources are reported by the chkdsk command on a PC
or by the top command on a UNIX workstation.
Do not forget that ICM itself will additionally allocate some
BufferSpace specified in the icm.cfg file.
Default (10.0) Mb
regularization procedure).
Default (30).
- find molecule - chemical substructure search
- find pattern - find sequence pattern in sequences of mol. objects.
- find database - advanced sequence similarity search
- align ms_1 ms_2 - alternatives solutions for 3D superposition
- find profile - find protein Prosite profiles in a sequence
- find prosite - find protein Prosite patterns in a sequence
Table( profile P_seqArray [ threshold ] )
Default (500).
minimize,
and as a part of one step of a multistep procedure in
montecarlo,
ssearch,
convert .
The number of function evaluations required to find the local minimum varies
widely depending on the terms used (i.e. the "tz" term makes
minimization very slow, if structure is far from its target).
If the minimum is found according to the
tolGrad
criterion, the procedure will be terminated anyway.
Default (100).
See also:
minimizeMethod , tolGrad , drop .
montecarlo command.
Since the procedure performs random steps
accompanied by local minimization (controlled by the mncalls parameter),
the number of function evaluations for the whole procedure can
be roughly evaluated as a product of mncalls
and the number of MC iterations. mncallsMC should be sufficiently
large to ensure convergence of the global optimization procedure and
may range from 10,000 for a single side-chain, 100,000 for a 3-4 residue
peptide to several million calls for 15-20 residue peptide or a large
protein loop.
Default (
1000 ).
The default value is small to minimize damage of the unintentional
calls of the montecarlo command.
See also:
montecarlo , mncalls .
stack .
The stack stops growing after this number is achieved
and starts replacing representative conformations with higher energy values
by new conformations with superior energies, if the latter are found.
Default (50)
See also:
montecarlo , ssearch .
stack
because their energies are higher than energies of the stack conformations.
Therefore, the
montecarlo
procedure is walking in the high energy area and is probably wasting its time.
When this threshold is reached the procedure acts according to the
highEnergyAction
parameter.
Default (50)
See also:
mnvisits , mnreject , stack .
montecarlo
procedure.
When this threshold is reached the procedure acts according to the
rejectAction
parameter (which usually increases the simulation temperature).
Default (10)
See also:
mnvisits , mnhighEnergy .
stack
in the course of a
montecarlo
procedure.
When this threshold is reached the MC procedure acts according to the
visitsAction
parameter.
A visit is an event when a newly generated conformation finds
a slot with a similar conformation in it, but the stack conformation
is not replaced by the new one because it has a better energy.
The optimal mnvisits parameter grows with the size of the problem (it may be several hundred for a 15-20 residue peptide).
Default (50)
See also:
mnreject , mnhighEnergy .
Default (10), minimal number (8).
See also: montecarlo local.
ssearch . Normally
the whole [-180., 180.] range is divided into nSsearchStep parts.
In the local mode
(i.e. the search is performed around a particular conformation)
the total search range around each variable is defined by
the ssearchStep parameter (30. deg. by default)
Default (3) .
This variable can be used as a hint to run an ICM command in parallel if there multiple threads can be submitted. Here is a subset of the ICM commands and functions supporting this parallelization:
- add column function= FUNCTION(X_chems) .. # for most chemical property functions
- make flat X_chems rotate
- set charge formal auto
| (7.0) find tablecommandsenumerate tautomermake reactionlearn type="nn"ScoreDistance chemicalFind chemical- they are also used in scripts: _ridge, _ride , _combiRidge
Example:
read mol table "big.sdf" name="t"
set charge formal auto nProc=8 # run in 8 threads
See also: fork , Nof( fork ), wait , l_out (defines the parent), Index(fork [system|all] ) for shell-level parallelization of icm scripts
montecarlo , randomize , Random function.
Helpful if you need to reproduce exactly a calculation
which uses random number(s). If the variable has its zero default value,
the random function is seeded from the current time plus the process id.
Otherwise, if you explicitly redefine it before, let us say, a montecarlo run,
it will use the specified number.
Note that the randomSeed parameter can be set only once in the very beginning
of the session. If you redefine its value in the middle of the session, it will
not be used. To push the new value of the seed, use the set randomize i_newRandomSeed
command.
Default (0).
Examples:
randomSeed=2493059372 # this number you took from the previous run
montecarlo # simulation will reproduce the previous one
#
...
#
set randomize 2493059372
montecarlo
Default (3).
Default (10).
See also:
sequenceLine .
The same parameter also controls the size of alignment block as saved by the write alignment command.
Example:
read alignment s_icmhome+"sh3"
sequenceLine=1000
write sh3 "aaa"
Default (60). Values >= 1 .
See also:
sequenceBlock
surfaceAccuracy to 5 in the rebel and make boundary calculations.
The corresponding number of dots per sphere is the following:
- Level 1 ( 89 dots )
- Level 2 ( 144 dots )
- Level 3 ( 233 dots )
- Level 4 ( 377 dots )
- Level 5 ( 610 dots )
Default (3)
See also: show area surface, "sf" energy term .
Smooth
function.
Default (7).