INDOCK for DOCK 3.6

What follows is a documented sample INDOCK file for DOCK 3.6. Many lines are required, lines starting with # are comments.

Required first line:

DOCK 3.5 parameter
###############################################################################
################## DOCK 3.5 INPUT PARAMETERS 2011/10/26 #######################
###############################################################################
###############################################################################
#                          INPUT/OUTPUT
#

This is the path to the receptor matching spheres file. Most scripts make a set of directories and copy the INDOCK file into them, so this path sometimes has an extra set of "../" in it compared to what you might think. If you use DOCK Blaster. Generally, match3 has more spheres than match2, so produces more possible orientations. These spheres are matched to ligand spheres, generated from heavy atoms in the "rigid component" of each ligand. For more about the rigid component, see Flexibase Format.

receptor_sphere_file          ../../sph/match2.sph

The next line is always 1, and is marked for deprecation.

cluster_numbers               1

The next line refers to which ligand file to use. If using many of the automated scripts, split_database_index is used, as this allows many ligand files (or just 1) to be placed in the split_database_index file and read in one after another during a DOCK run. If docking small things on your own, you can change this to any file.

# NOTE: split_database_index is reserved to specify a list of files
ligand_atom_file              split_database_index

This will control the file output, again many of the automated scripts expect it to be test. OUTDOCK files are always named OUTDOCK.

output_file_prefix            test.

This controls the random seed used in the minimization procedure. Changing this will produce slightly different results.

random_seed                   777
#
###############################################################################
#                             MATCHING
#

distance_tolerance is how different the distances can be between a pair of receptor matching spheres and a pair of ligand matching spheres for them to still be considered matched.

distance_tolerance            1.5

This changes how many spheres must be matched to generate an orientation. 3 as a minimum, 4 as a maximum is generally accepted as the right thing to use. Less than 3 is too degenerate to generate an actual orientation, and requiring more than 4 matched spheres does not work well, since we only use heavy atoms in ring systems to generate ligand matching spheres.

nodes_maximum                 4
nodes_minimum                 3

The next 4 parameters control how the histograms of distance differences are generated. The binsize is how big the bins are, the overlap controls if a sphere can be put into multiple bins. The ligand & receptor parameters are not required to be the same.

ligand_binsize                0.4
ligand_overlap                0.2
receptor_binsize              0.4
receptor_overlap              0.2

Bumping is using a quick check of distances when placing ligand atoms in the binding site to determine if they have a steric clash. The maximum is how many can be 'bumped' or in close steric contact per rigid or flexible component of the ligand, as per the Flexibase Format. Even ligands with some steric clashes can sometimes be rescued by minimization. Setting this number very high will cause many clashed orientations to be scored, which can be prohibitively slow.

bump_maximum                  1

The next four parameters are unused and unsupported.

focus_cycles                  0
focus_bump                    0 
focus_type                    energy
critical_clusters             no
#
###############################################################################
#                             COLORING
#

This controls whether chemical matching or coloring is used at all. If yes, many match lines are necessary. These may not be perfect, but DOCK Blaster has been using these for a long time. Setting this to no produces many more matched orientations, which can be slow, but can help you understand exactly what the energy function is doing.

chemical_matching             yes
case_sensitive                no
#                             ligand color, receptor color
match                         positive negative
match                         positive negative_or_acceptor
match                         positive not_neutral
match                         negative positive
match                         negative positive_or_donor
match                         negative not_neutral
match                         donor acceptor
match                         donor donacc
match                         donor negative_or_acceptor
match                         donor neutral_or_acceptor_or_donor
match                         donor not_neutral
match                         acceptor donor
match                         acceptor donacc
match                         acceptor positive_or_donor
match                         acceptor neutral_or_acceptor_or_donor
match                         acceptor not_neutral
match                         neutral neutral
match                         neutral neutral_or_acceptor_or_donor
match                         ester_o donor
match                         ester_o donacc
match                         ester_o positive_or_donor
match                         ester_o not_neutral
match                         amide_o donor
match                         amide_o donacc
match                         amide_o positive_or_donor
match                         amide_o not_neutral

Single mode is deprecated, these parameters won't work. See Dock Ligand Clustering

#
###############################################################################
#                             SINGLE MODE
#
#rmsd_override                 0.0
#contact_minimum               0
#energy_maximum               1.0e+6
##truncate_output              1000.0
#

Search mode is now the default/only mode of docking. Each parameter is described below.

###############################################################################
#                             SEARCH MODE
#

The ratio_minimum parameter has been slated for deprecation.

ratio_minimum                 0.0

These parameters control how many atoms are necessary in the ligand for it to be docked.

atom_minimum                  5 
atom_maximum                  100

How many of the top molecules will be saved in the output test.* file.

number_save                   50000

The maximum number of molecules that will be scored in any given run.

molecules_maximum             300000 

How many molecules will be skipped, this feature currently does not work.

initial_skip                  0

How long a molecule is processed before quitting. This feature currently may not work as expected.

timeout                       180

There are many scoring options:

# 
###############################################################################
#                             SCORING
#

Valid options for ligand_desolvation are 'volume' (partial desolvation a la Mysinger & Shoichet 2010), 'full' meaning that the entire ligand is assumed to be desolvated in the binding site and 'none', where no desolvation penalties are applied.

ligand_desolvation            volume

See the note about relative paths for the matching spheres above, the same comments apply here. There are 2 ways to run 'volume' or partial desolvation, one is to use one grid for every ligand atom like this:

solvmap_file                  ../../grids/solvmap_sev

The other option is to use one grid for ligand heavy atoms and one for ligand hydrogen atoms, you'll want to uncomment these lines to use them (and comment out the other solvmap_file line).

#solvmap_file                  ../../grids/solvmap_sev.heavy
#hydrogen_solvmap_file         ../../grids/solvmap.sev.hydrogen

This is the phimap file used for electrostatic scoring. For a better understanding of this grid, see Visualizing delphi. Sometimes this will change if you are using the new Qnifft Delphi maps, see Qnifft DOCK 3.6 conversion.

delphi_file                   ../../grids/rec+sph.phi

This controls the chemgrid file, which contains the van der Waals scoring for every coordinate (chem.vdw will be called) as well as the distance map grids that will be used for deciphering bumping (chem.bmp will be called).

chemgrid_file_prefix          ../../grids/chem

This is the parameter file that contains the atom type definitions:

vdw_parameter_file            ../../grids/vdw.parms.amb.mindock

The following options allow the electrostatics and van der Waals parameters to be scaled relative to each other and the solvation scoring.

electrostatic_scale           1.0
vdw_scale                     1.0

The following parameter lets ligands with internal steric clashes attempt to find a ligand conformation that scores well but does not have any internal clashes. Sometimes this procedure will fail in circumstances where there are many flexible branches, or where a ligand that is too large for the binding site is being docked.

check_clashes                 yes

If set to yes, this removes the positive solvation from each ligand atom and spreads it evenly over the molecule. This is deprecated because it does unexpected things to solvation, and will be removed entirely soon.

remove_positive_solvation     no

After each orientation of the rigid component is processed and the many ligand conformations have been examined, the best ligand conformation for that orientation can be minimized using the following parameters.

#
###############################################################################
#                             MINIMIZATION
#

No turns off minimization completely.

minimize                      yes

Don't minimize molecules that score above the minimization_max.

minimization_max             1.0e15

If set to yes, this checks to see if the orientation has already been scored and quits. This has not been tested recently.

check_degeneracy              no

How many iterations of minimization to do. More means longer run times, but potentially better poses.

simplex_iterations            250

How much the total energy can changed to be considered converged. Setting this higher will stop faster, setting it lower will cause it to do more iterations before converging (or potentially hitting the iteration max above).

simplex_convergence           0.1

If the energy changes by this much, restart the minimizer from this newest position.

simplex_restart               1.0

This is the initial distance in angstroms the molecule is translated (note that translation and rotation used to be swapped for many releases of DOCK).

simplex_initial_translation   0.2

How many degrees of initial rotation are done.

simplex_initial_rotation      5.0
#
###############################################################################
###############################################################################