usage: latticesim.pl [options] seqFile options: [-chain file | -rnd | -pdb file] [-g gridsize] [-par tsteps=val,ncycle=val,icycle=val, stiff=val,short=val,central=val,kdcore=val] [-l force min:max[=min:max ...]] [-d force res1:res2[=res1:res2 ...]] [-sa temp] [-const temp] [-norun]
This script provides an interface for running
low resolution lattice-based protein simulations with MONSSTER.
It creates all necessary input files and runs the MONSSTER program
mfold in the current directory.
The only required argument is a sequence file in MONSSTER format that can be generated with <docmark>genseq.pl</docmark>. By default or when -rnd is specified a random chain is generated as the starting configuration. The option -g can be used to provide the grid size used in the MONSSTER program. If this option is not given a default value of 100 is used. A specific starting configuration can be specified either directly with -chain in MONSSTER chain format or by providing a PDB file with -pdb that is used to generate a chain.
For fragment/loop modeling the option -l can be used to specifiy ranges of residues that are simulated fully flexible while the other surrounding residues are restrained according to the force constant specified. See <docmark>genrestr.pl</docmark> for an explanation how the restraint list is generated. If -l is used along with the -pdb option the input chain is generated from the PDB file only for residues outside the list in the argument of -l with random positions for the flexible fragment residues. During fragment/loop modeling a complete starting conformation for all residues can only be provided in form of a MONSSTER chain file by using the -chain option.
The option -d can be used to provide residue-residue distance restraints that may be available from known disulfide linkages or other information about residue-residue contacts. This option is particularly useful for structure predictions starting from random configurations by limiting the search space significantly.
Two main modes are available for lattice simulations: Simulated annealing and constant temperature. During simulated annealing the temperature is decreased continuously at each temperature step from an initially high value to room temperature. MONSSTER uses an intrinsic temperature scale where 1.0 corresponds approximately to room temperature. Typical simulated annealing runs start at temperatures between 2.0 and 2.5. The option -sa is used to select simulated annealing and provide the initial temperature. Constant temperature simulations are selected with -const followed by the desired temperature in the same units. This option also changes default values for the number of temperature steps to 1 and potential scaling of the centrosymmetric potential to 0.5. By default simulated annealing starting at 2.5 is performed.
Additional options can be specified with -par as a list of hash-type key/value pairs. A number of parameters are available to specify the length of simulation cycles:
tsteps selects the number of temperature steps. This affects the rate of cooling during simulated annealing but may be set to 1 for constant temperature simulations. The default value is 10. ncycle specifies the number of cycles at each temperature step (20 by default) and icycle specifies the number of Monte Carlo moves per cycle (50 by default).
For selective scaling of specific terms in the force field the following options (specified with -par) are available:
central sets scaling of the centrosymmetric potential (default: 0.25) that is used to favor more compact structures over extended forms. This is helpful if simulations are started from random chains but may be reduced for simulations of already compact conformations to avoid a bias towards spherical and too compact structures.
The parameter stiff controls the scaling of a generic potential term (default: 1.25) favoring the formation of regular protein secondary structure elements. For loop modeling reduction of this potential term may be beneficial.
short determines scaling of short range interactions (default: 0.5) that depend on the type of secondary structure elements provided in the sequence file. Increasing this potential term may be useful to increase the bias towards known secondary structure.
kdcore controls the rate of hydrophobic collapse. The default value of 0.9 results in relatively rapid collapse to compact structures with hydrophobic cores. Reduction of this value up to 0.1 results in more realistic sampling of the folding landscape but with a lower yield of near-native structures in simulated annealing runs.
Finally, the -norun option can be given to request only a setup of all necessary files to run MONSSTER without actually calling the program to start the simulation.
- usage information
runs a simulated annealing lattice simulation starting from a random configuration with default simulation parameters
latticesim.pl -l 50.0 10:21 -pdb 1vii.exp.pdb -const 1.0 -par ncycle=100 1vii.seq
runs a constant temperature lattice simulation at approximately room temperature (1.0 in MONSSTER temperature units). The initial chain is generated from 1vii.exp.pdb with random positions for residues 10 through 21. During the simulation residues below 10 and above 21 are restrained with a variable force constant up to 50.0 (in internal MONSSTER energy units). The -const option for constant temperature automatically selects a single temperature step. Finally -opt ncycle=100 increases the number of cycles to 100.
latticesim.pl -sa 2.0 -d 5.0 10:34=19:26 -chain 1vii.exp.chain 1vii.seq
runs a simulated annealing lattice simulation starting at a temperature of 2.0 (internal MONSSTER units). The input chain is taken from a chain file. Distance restraints with a force constant of 5.0 (internal units) are set between residues 10 and 34 and 19 and 26.