Personal tools
You are here: Home FAQs & Help FAQs
Document Actions

FAQs

View entire FAQ in full Up to table of contents
Frequently Asked Questions (& Answers!)

Scientific Questions

Can AutoDock be used for "Blind Docking"?

Can AutoDock be used when the structure of the ligand and the protein are both known, but the location of the binding site is unknown?

Yes, AutoDock can be used when the location of the binding site is unknown. This is often referred to as "blind docking", when all that is known is the structure of the ligand and the macromolecule.

It will be necessary to set up the dockings to search the entire surface of the protein (or other macromolecule) of interest. This can be achieved using AutoGrid to create very large grid maps, with the maximum number of points in each dimension, and if necessary, creating sets of adjacent grid map volumes that cover the macromolecule. The third-party tool BDT can be used to set up such sets of grid maps.

Several authors have used AutoDock to perform blind docking (see 1.-6.); for instance, Hetenyi et al. published two papers showing that AutoDock can be used to perform blind docking of peptides to proteins, and drug-sized molecules to proteins.

  1. Hetenyi, C. and van der Spoel, D. (2002) Efficient docking of peptides to proteins without prior knowledge of the binding site. Protein Science, 11(7): 1729-1737.
  2. Kovacs, M., Toth, J., Hetenyi, C., Malnasi-Csizmadia, A., and Sellers, J.R. (2004) Mechanism of blebbistatin inhibition of myosin II. Journal of Biological Chemistry, 279(34): 35557-35563.
  3. Bikadi, Z., Hazai, E., Zsila, F., and Lockwood, S.F. (2006) Molecular modeling of non-covalent binding of homochiral (3S,3 ' S)-astaxanthin to matrix metalloproteinase-13 (MMP-13). Bioorganic & Medicinal Chemistry, 14(16): 5451-5458.
  4. Hazai, E., Bikadi, Z., Zsila, F., and Lockwood, S.F. (2006) Molecular modeling of the non-covalent binding of the dietary tomato carotenoids lycopene and lycophyll, and selected oxidative metabolites with 5-lipoxygenase. Bioorganic & Medicinal Chemistry, 14(20): 6859-6867.
  5. Hetenyi, C. and van der Spoel, D. (2006) Blind docking of drug-sized compounds to proteins with up to a thousand residues. FEBS Letters, 580(5): 1447-1450.
  6. Iorga, B., Herlem, D., Barre, E., and Guillou, C. (2006) Acetylcholine nicotinic receptors: finding the putative binding site of allosteric modulators using the "blind docking" approach. Journal of Molecular Modeling, 12(3): 366-372.

This FAQ applies to: AutoDock 3, AutoDock 4

Where do I set the AutoDock 4 force field parameters?

Where can I see or change the values of the van der Waals parameters, hydrogen bonding parameters, and/or atomic solvation parameters ? Where can I see which AutoDock 4 atom types are supported, and what parameters they correspond to? Where can see or modify the values of the linear regression coefficients for the linear free energy model?

In AutoDock 4, we have introduced a new command "parameter_file" that takes a new parameter library file that contains the various force field parameters.  In most cases, you will not need to modify these values, but it is important to know where they are and how to change them if necessary.

The standard AutoDock 4 parameters are in the file "AD4_parameters.dat" which can be found in the "autodocksuite-4.n.m/src/autodock-4.x.y" directory of the AutoDock 4 distribution, where n and m are the major and minor version numbers of the AutoDock Suite, and x and y are the major and minor version numbers of AutoDock.  Here is an example:

# $Id: AD4_parameters.dat,v 1.14 2007/04/27 06:01:47 garrett Exp $
# 
# AutoDock 
# 
# Copyright (C) 1989-2007,  Garrett M. Morris, David S. Goodsell, Ruth Huey, Arthur J. Olson, 
# All Rights Reserved.
# 
# AutoDock is a Trade Mark of The Scripps Research Institute.
# 
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
# 
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
# 
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.

# AutoDock Linear Free Energy Model Coefficients and Energetic Parameters
#                        Version 1.0
#                     $Revision: 1.14 $

# AutoDock 4 free energy coefficients with respect to original (AD2) energetic parameters
#
#               Free Energy Coefficient
#               ------
FE_coeff_vdW    0.1560
FE_coeff_hbond  0.0974
FE_coeff_estat  0.1465
FE_coeff_desolv 0.1159
FE_coeff_tors   0.2744

# AutoDock 4 Energy Parameters

# - Atomic solvation volumes and parameters
# - Unweighted vdW and Unweighted H-bond Well Depths
#
# - Atom Types
# - Rii = sum of vdW radii of two like atoms (in Angstrom)
# - epsii = vdW well depth (in Kcal/mol)
# - vol = atomic solvation volume (in Angstrom^3)
# - solpar = atomic solvation parameter
# - Rij_hb = H-bond radius of the heteroatom in contact with a hydrogen (in Angstrom)
# - epsij_hb = well depth of H-bond (in Kcal/mol)
# - hbond = integer indicating type of H-bonding atom (0=no H-bond)
# - rec_index = initialised to -1, but later on holds count of how many of this atom type are in receptor
# - map_index = initialised to -1, but later on holds the index of the AutoGrid map
# - bond_index = used in AutoDock to detect bonds; see "mdist.h", enum {C,N,O,H,XX,P,S}
#
# - To obtain the Rij value for non H-bonding atoms, calculate the 
#        arithmetic mean of the Rii values for the two atom types.
#        Rij = (Rii + Rjj) / 2
#
# - To obtain the epsij value for non H-bonding atoms, calculate the 
#        geometric mean of the epsii values for the two atom types.
#        epsij = sqrt( epsii * epsjj )
#
# - Note that the Rij_hb value is non-zero for heteroatoms only, and zero for H atoms;
#        to obtain the length of an H-bond, look up Rij_hb for the heteroatom only;
#        this is combined with the Rii value for H in the receptor, in AutoGrid.
#        For example, the Rij_hb for OA-HD H-bonds will be (1.9 + 1.0) Angstrom, 
#        and the weighted epsij_hb will be 5.0 kcal/mol * FE_coeff_hbond.
#
#        Atom   Rii                             Rij_hb       rec_index
#        Type         epsii           solpar         epsij_hb    map_index
#                            vol                          hbond     bond_index
#        --     ----  -----  -------  --------  ---  ---  -  --  -- --
atom_par H      2.00  0.020   0.0000   0.00051  0.0  0.0  0  -1  -1  3    # Non H-bonding Hydrogen
atom_par HD     2.00  0.020   0.0000   0.00051  0.0  0.0  2  -1  -1  3    # Donor 1 H-bond Hydrogen
atom_par HS     2.00  0.020   0.0000   0.00051  0.0  0.0  1  -1  -1  3    # Donor S Spherical Hydrogen
atom_par C      4.00  0.150  33.5103  -0.00143  0.0  0.0  0  -1  -1  0    # Non H-bonding Aliphatic Carbon
atom_par A      4.00  0.150  33.5103  -0.00052  0.0  0.0  0  -1  -1  0    # Non H-bonding Aromatic Carbon
atom_par N      3.50  0.160  22.4493  -0.00162  0.0  0.0  0  -1  -1  1    # Non H-bonding Nitrogen
atom_par NA     3.50  0.160  22.4493  -0.00162  1.9  5.0  4  -1  -1  1    # Acceptor 1 H-bond Nitrogen
atom_par NS     3.50  0.160  22.4493  -0.00162  1.9  5.0  3  -1  -1  1    # Acceptor S Spherical Nitrogen
atom_par OA     3.20  0.200  17.1573  -0.00251  1.9  5.0  5  -1  -1  2    # Acceptor 2 H-bonds Oxygen
atom_par OS     3.20  0.200  17.1573  -0.00251  1.9  5.0  3  -1  -1  2    # Acceptor S Spherical Oxygen
atom_par F      3.09  0.080  15.4480  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Fluorine
atom_par Mg     1.30  0.875   1.5600  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Magnesium
atom_par MG     1.30  0.875   1.5600  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Magnesium
atom_par P      4.20  0.200  38.7924  -0.00110  0.0  0.0  0  -1  -1  5    # Non H-bonding Phosphorus
atom_par SA     4.00  0.200  33.5103  -0.00214  2.5  1.0  5  -1  -1  6    # Acceptor 2 H-bonds Sulphur
atom_par S      4.00  0.200  33.5103  -0.00214  0.0  0.0  0  -1  -1  6    # Non H-bonding Sulphur
atom_par Cl     4.09  0.276  35.8235  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Chlorine
atom_par CL     4.09  0.276  35.8235  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Chlorine
atom_par Ca     1.98  0.550   2.7700  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Calcium
atom_par CA     1.98  0.550   2.7700  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Calcium
atom_par Mn     1.30  0.875   2.1400  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Manganese
atom_par MN     1.30  0.875   2.1400  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Manganese
atom_par Fe     1.30  0.010   1.8400  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Iron
atom_par FE     1.30  0.010   1.8400  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Iron
atom_par Zn     1.48  0.550   1.7000  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Zinc
atom_par ZN     1.48  0.550   1.7000  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Zinc
atom_par Br     4.33  0.389  42.5661  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Bromine
atom_par BR     4.33  0.389  42.5661  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Bromine
atom_par I      4.72  0.550  55.0585  -0.00110  0.0  0.0  0  -1  -1  4    # Non H-bonding Iodine


These are the default values that are "baked in" to AutoGrid 4 and AutoDock 4, and are the values that will be used if you do not specify your own parameter file explicitly in your GPF or DPF with the "parameter_file" command.

You can add new atom types and parameters to this file, or modify the existing ones.  If you come up with new or modified parameters, we would like to know about them, so we can incorporate them in future releases.  Thanks!

This FAQ applies to: AutoDock 4

Should I always use polar hydrogens?

I know that AutoDock needs hydrogen atoms on the macromolecule and the ligand, but should I use polar hydrogens?

Yes, for both the macromolecule and the ligand, if you want to use the AutoDock 4 force field properly, you should always add all hydrogens, compute Gasteiger charges and then  merge the non-polar hydrogens.  This is because AutoDock 4 uses the united atom model to represent molecules, and the AD4 scoring function was calibrated using Gasteiger partial charges on both the ligand and the macromolecule.

Polar hydrogens are hydrogen atoms that are bonded to electronegative atoms like oxygen and nitrogen.  (ADT assumes that non-polar hydrogens are hydrogens bonded to carbon atoms.)

This FAQ applies to: AutoDock 3, AutoDock 4

How can I set up the protonation state of my histidine sidechains?

Histidines can be neutral or positively charged. When neutral, they can be protonated at the delta (HD1) or epsilon (HE2) positions. How can I set these up?

Using ADT


There is a command in ADT to help you decide on the protonation of the Histidines, but you have to load the commands before you can use it: go to "File > Load Module" and then scroll down, click on "repairCommands", and then click "Load Module" followed by "Dismiss". Now, go to "Edit > Hydrogens > Edit Histidine Hydrogens".

If there are any histidines in your molecule, a panel will open up listing each histidine residue along with a row of radio buttons. You can use these to choose whether each histidine should be neutral, HD1; neutral, HE2; or protonated.

Using Reduce/Molprobity


There is a very nice tool called Reduce (with a web-accessible front end called Molprobity) that can be used for adding hydrogens and optimising the hydrogen-bond network by flipping amido groups in Asn and Gln sidechains, and His imidazole rings by 180º. It can also be used for evaluating the quality of your protein structure. See:

Word, et al. (1999) "Asparagine and glutamine: using hydrogen atom contacts in the choice of sidechain amide orientation" J. Mol. Biol. 285, 1733-1745.

This FAQ applies to: AutoDock 3, AutoDock 4

How do I add new atom types to AutoDock 4?

AutoDock version 4.0 has parameters for H, C, N, O, F, Mg, P, S, Cl, Ca, Mn, Fe, Zn, Br and I. What do you do if your molecule has an atom type that isn't already parameterised?

The various parameters in the AutoDock 4 scoring function are described in another FAQ, Where do I set the AutoDock 4 force field parameters?. The atom types' names and parameters are specified in a file that can be called anything, but by default is called "AD4_parameters.dat". You can find a copy of this default file in the source code of AutoGrid and AutoDock. This parameter file can be specified by the "parameter_file" keyword in the GPF and DPF, but if this keyword is not given, then AutoGrid and AutoDock use the default values. So the only time you need to use the "parameter_file" keyword is when you want to change the default values, or to add new atom types.

How do I add new atom types?

Important: Bear in mind that the AutoDock 4 scoring function was calibrated for the current set of atom types, and that if you add new ones, strictly speaking you should perform a re-calibration of the force field to determine the correct coefficients for the molecular mechanics terms and the empirical term of the linear free energy model.

You will need to find, compute or set the following values for each new atom type, which you specify after the "atom_par" keyword. Note that the values on each "atom_par" line are space delimited, not fixed width. You will need to add one "atom_par" line for every new atom type. It is possible to define 'synonyms' for atom types, by repeating the numerical atom parameter lines for every variant of the atom type's one- or two-character name.

  • name of the atom type; this can be one or two characters long, and should correspond to the atom type at the end of ATOM or HETATM lines in PDBQT files; specify this in the "Atom Type" field
  • van der Waals radius of the atom/ion (in Angstrom); specify twice this value in the "Rii" field
  • epsilon or energy well depth for two like interacting atoms/ions (in Kcal/mol); specify in the "epsii" field
  • volume of the atom/ion (in Angstrom^3); compute this from 4/3 * PI * (Rii/2)^3; specify in the "vol" field
  • atomic solvation parameter of the atom/ion; the ai or "ASP" values in the equation for Si in the description of the Desolvation Free Energy Term in AutoDock 4; specify this in the "solpar" field
  • hydrogen bonding radius (in Angstrom); for non-hydrogen-bonding atom types this is 0.0; specify this in the "Rij_hb" field
  • hydrogen bonding energy well depth (in Kcal/mol); for non-hydrogen-bonding atom types this is 0.0; specify this in the "epsij_hb" field
  • hydrogen bonding type; an integer; for non-hydrogen-bonding atom types this is 0; specify this in the "hbond" field
  • receptor type index; an integer; default value is -1; specify this in the "rec_index" field
  • grid map index; an integer; default value is -1; specify this in the "map_index" field
  • bond type index; an integer; default value is 4; specify this in the "bond_index" field

The last six values (0.0 0.0 0 -1 -1 4) will be the same if the atom type is not involved in hydrogen bonding.

Note that a comment can be specified after a number sign symbol, #

Remember to use the "parameter_file AD4_parameters.dat" command in both the GPF and DPF, so that the new parameters are used in both the AutoGrid pre-calculation of the grid maps, and in the AutoDock dockings. Specify this command on the first line of the GPF and DPF files.

This FAQ applies to: AutoDock 4

How Autodock 4 converts binding energy (kcal/mol) into Ki

Shows formula Autodock is using to convert binding energy into Ki.

This is taken from printEnergies.cc that is distributed with AutoDock source code:

    // equilibrium:   E  +  I  <=>    EI
    // binding:       E  +  I   ->    EI         K(binding),      Kb
    // dissociation:     EI     ->  E  +  I      K(dissociation), Kd
    //
    //                            1
    //         K(binding) = ---------------
    //                      K(dissociation)
    // so:
    //      ln K(binding) = -ln K(dissociation)
    //              ln Kb = -ln Kd
    // Ki = dissociation constant of the enzyme-inhibitor complex = Kd
    //      [E][I]
    // Ki = ------
    //       [EI]
    // so:
    //              ln Kb = -ln Ki
    // deltaG(binding)    = -R*T*ln Kb
    // deltaG(inhibition) =  R*T*ln Ki
    //
    // Binding and Inhibition occur in opposite directions, so we
    // lose the minus-sign:  deltaG = R*T*lnKi,  _not_ -R*T*lnKi
    // => deltaG/(R*T) = lnKi
    // => Ki = exp(deltaG/(R*T))


See also: ADL: calculation of Ki in Autodock 4

This FAQ applies to: AutoDock 4

Missing atom types (“Why atom type 'X' is not recognized?”)

AutoDock has been parameterized to support atom types that are most frequently encountered in biological systems, and were included in the complexes used to calibrate the force field. For this reason, many of the less common atom types are not included in the force field parameters. There are several ways of dealing with this problem, depending on the trade between the level of complexity involved in modeling them and accuracy obtained: The rigorous way. Find appropriate parameters for the new atom type and add them to the existing forcefield parameter file. AutoGrid and AutoDock may then be run using parameters from the file instead of the internal parameters. The format for the parameter file is described in full in the AutoDock Manual. This process can be simplified by modifying one of the existing atom types and simply changing the vdW radii to give an approximate set of parameters. The easy way. Substitute the atom type with the closest type already present in the existing forcefield (for instance, using C or A for boron). This change is made by manually changing the PDBQT file, substituting the atom type of the atom with an existing atom type. The atomic charge may also be set manually.

The next AutoDock release will include extended atom types and it will simplify the process of adding missing atom types.

Water: my target crystal structure has some waters in the active site, what should I do with them?

Often, inclusion of an important water can improve the docked results. We have also developed a new method for including multiple sites of hydration--please see (reference) for more information. Some of the automated scripts will remove water and other non-protein components, but you can use the GUI to process coordinates to create a PDBQT file. The coordinates are input, hydrogens added, and charges calculated using the menus in the PMV portion of the GUI. This process is described in more detail in the tutorial. Note that the current Babel-based method for adding hydrogens will not optimize the position of the hydrogens--it is often better to add these hydrogens manually using an external molecular modeling program. Alternatively, a new method has been developed for predicting position of waters and their displacement with AutoDock 4.x (Forli and Olson 2012, see User Manual referrences)

Side Chain Alternate Confirmations

Protein flexibility is one of the major challenges that will cause difficulty with docking results. If the protein has alternate conformations for sidechains, the best option is to edit the files to great several receptor files, each with a different one of the alternatives. The automated scripts for receptor preparation will typically save only conformation A.

Co-factors and Nucleic Acids

Some of the automated scripts will remove non-protein components, but you can use the GUI to process coordinates to create a PDBQT file. The coordinates are input, hydrogens added, and charges calculated using the menus in the PMV portion of the GUI. This process is described in more detail in the tutorial. Note that the current Babel-based method for adding hydrogens will not optimize the position of the hydrogens and often makes mistakes in hydrogen placement for non-protein molecules--it is often better to add these hydrogens manually using an external molecular modeling program.

Metal Charges

The current Babel-based method for preparation of coordinates does not handle metal charges. These may be assigned manually by editing the PDBQT file. Note, formal charges may be too strong for the current force field parameterization.

Lowest energy or Largest Cluster? How to evaluate docking results

Researchers in our laboratory have had good luck using both of these metrics in virtual screening, but we have not yet done a comprehensive study to show the effectiveness of one over the other. Our best results have been obtained by using one of these metrics and combining it with a biological metric, such as the proximity to a group on the receptor which is known to be important in the interaction (if this type of information is known). See Expert Opinion in Drug Discovery (2010) 5, 597-607 for a description of metrics that have been effective.

by morris last modified 2007-07-19 17:31
Contributors: Ruth Huey, Garrett M. Morris, Sargid Dallakyan, Stefano Forli

Powered by Plone CMS, the Open Source Content Management System

This site conforms to the following standards: