Home Linus Helper Programs Examples File Descriptions Index

Default Parameters

 


LINUS

 

The default parameters used by LINUS may be changed by including one of the following entries in the LINUS run file linus.py. Instructions for overriding the parameters listed below are described in the Sample LINUS Run Files section.

The default simulation parameter values are:

beta= 2.0, beta = 1/RT. A value of 2.0 approximates 25 deg C in the context of LINUS.

> numcycle= 1000, A cycle is N-2 moves where N is the number of residues in the polypeptide. A move consists of choosing, at random, one of four secondary structure phi, psi combinations and applying these to residue i, i-1 and i+1. Residue i may be any residue 2 to N-1 and is chosen randomly for each move.

numsave= 10, Save a structure every 10 cycles.

trials= 1, Number of simulation trials to run. See next two parameters which relate to this.

updateweights= 0, Count fraction of different secondary structure fractions at end of trial and weight choice of secondary structures for next trial according to fraction; yes, 1; no, 0

reset_conformation= 0, Reset conformation at end of trial back to original input conformation (usually extended); yes, 1; no, 0

write_rc= 0, Write out a file containing a mesostate list for each accepted conformation (this can be large); yes, 1; no, 0

updatepmeso= 0, Calculate new probability mesostates at the end of each cycle; yes, 1; no, 0

reset_emin_struc= 0, Pass the current cycle minimum energy conformation to the next cycle rather than the conformation in memory at the end of the cycle; yes, 1; no, 0

use_lock_ss= 0, After each move test for helix or beta-hairpin and take residues in these structures out of the movelist 90 percent of the time; yes, 1; no, 0

use_msd_sampling= 0, [NOTE: Do not use - calculation of mean is incorrect.]
At the end of each cycle calculate the mean and standard deviation of phi and psi for each residue; yes, 1; no, 0
If the standard deviation is < 5.0 (i.e. the residue seems to favor a small range of values) moves in the next cycle will be chosen from the range mean +/- standard deviation. This is an attempt to "focus" the choices of phi and psi when LINUS prefers a narrow range of values for a particular residue. If the standard deviation never goes below 5.0 this constraint will not come into play.

use_ms_sampling= 0, Choose phi and psi values from a mesostate bin listed in an input file called "mesostates"; yes, 1; no, 0
See here for a description of how to do this.

use_pms_sampling= 0, Choose phi and psi values from a probability distribution of mesostates; yes, 1; no, 0
See here for a description of how to do this.

use_local_mov= 0, A non-pivot move in the sense that a combination of backbone torsion angles for residues i-1, i, and i+1 are chosen so that downstream residues do not move (more specifically so that the CA atoms of residues i+3, i+4, i+5 and i+6 have an rmsd of less than 0.0001); yes, 1; no, 0
This doesn't do much to a strand conformation because it is hard to find three sets of phi/psi that do not move the downstream residues; but for turn conformations it allows a search of local conformational space.

use_dist_constraints= 0, Apply atom to atom distance restraints as part of the scoring function; yes, 1; no, 0
A set of pairwise atom distance restraints in the format of an XPLOR/CNS NOE file is read as input; then the score is sqrt(dist - dmax) if the dist > dmax, sqrt(dmin - dist) if the dist < dmin, and zero otherwise. See for an example.

The default H-bond parameters are:

use_hbond= 0, Use backbone to backbone hydrogen bonds: yes, 1; no, 0

use_sidechain_hbond= 0, Use sidechain to backbone hydrogen bonds: yes, 1; no, 0

hbond_distance= 3.0, Limit distance for H-bond score to be its maximum value (as defined in hbond_score_short).

hbond_probe= 1.5, Score scales to 0 at distance hbond_distance + hbond_probe (3.5 + 1.5 = 5.0).

hbond_score_short= 0.4, Score at <= hbond_distance distance for H-bonds to residues with sequence separation <= 6 residues.

hbond_score_long= 1.0, Score at <= hbond_distance distance for H-bonds to residues with sequence separation > 6 residues. There is no scaling for these H-bonds, the score goes to zero if the distance is larger than hbond_distance.

hbond_torsion= 130.0, Out of plane dihedral Oj - Ni - CAi - Ci-1 minimum (absolute value) for H-bond to form. This insures that N-H...O hydrogen bonds are +/- 40 degrees from directly in line with the N-H vector.

sidechain_hbond_distance= 3.0, Limit distance for side chain donor to back bone acceptor. There is no switching for these H-bonds, the score is zero beyond this limit.

sidechain_hbond_score= 1.0, As above for side chain donor to back bone acceptor.

sidechain_hbond_torsion= 140.0, As above.

hbond_winmin= 2, Minimum sequence separation between two residues that can participate in backbone/sidechain acceptor to backbone donor hydrogen bonds.

hbond_winmax= 6, Maximum sequence separation between two residues that can participate in backbone/sidechain acceptor to backbone donor hydrogen bonds. Insures local interactions only.

The default contact parameters are:

use_contact= 0, Use hydrophobic contacts; yes, 1; no, 0.

contact_probe= 2.8, Contact scored only if distance between the surfaces of designated atoms for each of the two residues is <= contact_probe distance.

contact_winmin= 2, Minimum sequence separation between two residues that can participate in contact energy calculations.

contact_winmax= 6, Maximum sequence separation between two residues that can participate in contact energy calculations.

contact_scale= 1.0, Multiplier of the contact energy term.

The default conformational weights are:

"helix", 0.15,"all",
"strand", 0125,"all",
"turn1", 0.225,"all",
"turn2", 0.225,"all",
"PII", 0.10,"all",
"coil", 0.15,"all",

Set sampling weights for specified move type and specified list of residues. These five values define the "smart move set" for Monte Carlo trial moves. The last argument is a list of residues; it could be "all" for all residues, "1, 5, 9" for residues 1, 5, and 9, or " 'gly', 'pro' " for all gly and pro residues (notice single quotes). See the Examples section for an example of a LINUS run file with a random (non-smart) move set.