9. MINIMIZATION PARAMETERS

The parameters that control the way that LIGPLOT minimizes a given plot (in terms of reducing the numbers of atom-atom, bond-bond and atom-bond overlaps) are given at the end of the ligplot.prm file.

It is unlikely you will need to alter these values, although sometimes you might get a better plot by altering the parameters by trial-and-error. The default parameters have been chosen to give reasonable results in most cases.

If the plot is very cluttered, you might improve it by increasing the "Stretch factor" parameter, described below.

Another way of improving a cluttered or unclear plot is by modifying the ligplot.pdb file, as described in section 6.

MINIMIZATION PARAMETERS
-----------------------
 10.00    <- Atom-atom clash parameter
  0.20    <- Bond-atom clash parameter
200.0     <- Bond-overlap score
  0.5     <- Weight for term giving H-bond deviation from ideal value
  0.01    <- Weight for term giving non-bond dist deviation from ideal value
 10.0     <- Weight for internal energies (atom clashes, etc)
 15.0     <- Furthest move-distance for H-bonded groups (in Angstroms)
  1.0     <- "Stretch factor" for H-bond lengths
1000      <- Number of loops for the minimization process
0.0       <- Terminate minimization if energy drops by less than this value
N         <- Random start for minimization routines - (Y/N)?
20.0      <- Weight for anchor-position energy term
5.0       <- Weight for interface-boundary energy term
2.0       <- Closest atom-distance to boundary representing interface
0.3       <- Weight for relative residue-positions energy term

Description of the parameters:-

Atom-atom clash parameter. The first parameter defines how to penalize clashes between atoms of the different objects on the plot, or indeed between atoms of the same object. The higher the value the more strongly are the atoms repelled from overlapping one another. The repulsion force itself is based on the inverse square of the distance between the outer edges of the atoms, within a preset interaction range. The parameter determines the weight to be assigned to this term in the final energy equation.

Bond-atom clash parameter. This parameter defines how to penalize any bond-atom overlaps. The higher the value the more strongly are bonds repelled from overlapping with atoms. As above the function follows an inverse square law, based in the closest distance between the bond and the edge of the atom.

Bond-overlap score. This is a single penalty on the overlap of any hydrogen bond with any other bond in the picture. The higher the value, the more are bond overlaps penalized.

Weight for term giving H-bond deviation from ideal value. This term determines how tightly the hydrogen bonds on the diagram are to be restrained to their true lengths. The higher the value, the tighter the restraint. (See also the "Stretch factor", below).

Weight for term giving non-bond dist deviation from ideal value. As for the H-bond deviation above, this term determines the weight applied to the restraints on the distances between atoms involved in hydrophobic contacts. The higher the value the more tightly are these distances restrained to their true values. (See also the "Stretch factor", below).

Weight for internal energies (atom clashes, etc). This is a factor by which the above energy parameters are to be upweighted when applied to calculating the self-interactions between atom-atom, atom-bond and bond-bond overlaps within a given object. For example, if this weight is set to 10.0, then, say, atom-atom clashes between atoms of the same residue will be punished 10 times more heavily than atom-atom clashes between atoms of this residue and those of other residues on the page.

Furthest move-distance for H-bonded groups (in Angstroms). This determines the maximum distance that any residue can be moved during the minimization process. If the plot is very complicated, with very many objects on it, you may need to increase this distance to allow residues a better chance of moving to more favourable locations.

"Stretch factor" for H-bond lengths. If the picture is very cluttered, with all the hydrogen-bonded groups clustering very tightly around the ligand, you can relax the tension a little by changing this parameter and allowing the hydrogen-bond and hydrophobic contact distances to be restrained to values greater than their actual values. So, by setting the stretch factor to 2.0, say, all these distances will be restrained to twice their actualy values and the residues on the plot will not pack quite so tightly around the ligand.

Number of loops for the minimization process. This parameter allows you to change the number of loops over which the minimization process runs. The default value is 1000.

Terminate minimization if energy drops by less than this value. This parameter allows you to terminate the energy minimization process before it reaches the number of loops specified above. If the energy drop between successive loops is smaller than the value specified here, the minimization will stop. It prevents the program cycling on and on for too long without making any significant change to the final diagram.

The default value is 0.0, which means that the minimization will run for the full number of cycles specified above.

Random start for minimization routines - (Y/N)? This parameter determines whether the random-number generator used in the energy minimization routines is to have the same start-point each time the program is run, or is to have a randomly-generated start-point.

If this option is set to "Y", a random seed for the next run of the program is written to the output file ranseed.dat. On the next run, this number is read in from the file and used as the seed for initializing the random number generator. This means that a slightly different LIGPLOT diagram might be generated for exactly the same structure each time the program is run, merely because the minimization proceeds down a slightly different pathway.

If the option is set to "N", the random number generator is started at the same place every time the program is run, and so will give identical results for the same structure (unless, of course, any of the other minimization parameters are altered).

Weight for anchor-position energy term. This parameter is used when applying positional restraints to the residues in the plot. The value determines the weight of these restraints. The restraints themselves are listed in the .rcm file and consist of centre-of-mass positions for some, or all, of the residues on the plot (see Improving the plots - Same protein different ligands). They are used when you want the disposition of the residues on the plot, around a particular ligand, to match the layout in a previous LIGPLOT produced for, say, a different ligand.

Weight for interface-boundary energy term. This parameter assigns the weight for the energy term measuring how far residues stray onto the wrong side of an interface boundary. Only used when plotting interactions across a dimer or domain-domain interface is being plotted (see DIMPLOT, section 8).

Closest atom-distance to boundary representing interface. This parameter gives the closest distance of approach to the interface boundary in DIMPLOT (see section 8) before a residue starts being penalised for straying onto the wrong side of the boundary.

Weight for relative residue-positions energy term. This parameter determines the weight of the energy term that measures how all the relative distances between residues on the plot deviate from the same distances in the original protein.