AutoDock Vina

The AutoDock Vina tool allows running ligand-receptor docking calculations with AutoDock Vina, using either a web service provided by the National Biomedical Computation Resource (NBCR) or a locally installed copy of the program. Docking results are shown automatically in ViewDock. See also: AddH, Dock Prep

The web service is a public (shared) resource, and this interface only allows docking a single ligand with very limited sampling. For a more intensive sampling of space, as needed for most research applications, or access to other options such as ligand database search, we recommend downloading the AutoDock Vina program and using it directly. Other limitations are given below.

Users should cite:

AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. Trott O, Olson AJ. J Comput Chem. 2010 Jan 30;31(2):455-61.

There are several ways to start AutoDock Vina, a tool in the Surface/Binding Analysis category. It is also implemented as the command vina.

The receptor and ligand structures should be opened as separate models in Chimera. If the receptor structure contains MSE (selenomethionine) residues, incomplete side chains, or atoms with alternate locations, running Dock Prep beforehand to correct those issues is recommended. The AutoDock Vina tool runs AutoDock accessory scripts locally to (further) prepare the structures, such as to add hydrogens if they have not been added already with Chimera.

Output file - location and filename prefix for output files. If the prefix is name, the files generated by a successful run will be:

name (or name.pdbqt if indicated in the file browser) - docking results in PDBQT format, automatically read into ViewDock when the calculation finishes
name.receptor.pdb - receptor PDB file from Chimera, input to the AutoDock receptor preparation script
name.receptor.pdbqt - processed receptor in PDBQT format, input to AutoDock Vina
name.ligand.pdb - ligand PDB file from Chimera, input to the AutoDock ligand preparation script
name.ligand.pdbqt - processed ligand in PDBQT format, input to AutoDock Vina
name.conf - AutoDock Vina configuration file

Receptor - the model to use as the receptor (choose from pulldown menu)

Ligand - the model to use as the ligand (choose from pulldown menu)

Receptor search volume options - definition of the box in which to sample ligand positions

Resize search volume using [button] - reassign the indicated mouse button to defining a search volume box. When there is no pre-existing box:
- Dragging with the assigned button sweeps out a box aligned with the X, Y, and Z axes, the smallest possible box containing all voxels under the initial and final cursor positions. Depending on the viewing angle, the box may be surprisingly large.
After a box has been created, it can be erased or further adjusted.
- Clicking the assigned button at a point not over the box erases the box.
- Dragging from a point over the box moves the frontmost face; also holding down Shift moves the the rearmost face instead of the frontmost.
- Dragging from a point not over the box translates the whole box in X and Y; Shift-dragging translates the box in the Z dimension.

Center: [x][y][z] - box center in the receptor coordinate system; can be edited directly
Size: [x][y][z] - box dimensions along X, Y, and Z in the receptor coordinate system; can be edited directly

Receptor options - settings for the receptor preparation script:

Add hydrogens in Chimera (true/false) - whether to add hydrogens in Chimera (see addh) before calling the script. The receptor prep script will check for hydrogens and add them if they are missing. AutoDock Vina needs the polar (potentially H-bonding) hydrogens to identify atom types for scoring purposes.
Merge charges and remove non-polar hydrogens (true/false) - note AutoDock Vina does not use charges or nonpolar hydrogens, so this setting is not expected to affect results except for the presence or absence of nonpolar hydrogens in the processed receptor
Merge charges and remove lone pairs (true/false) - note AutoDock Vina does not use charges or lone pairs, so this setting is not expected to affect results except for the presence or absence of lone pairs in the processed receptor (and there may not have been any lone pairs to start with)
Ignore waters (true/false)
Ignore chains of non-standard residues (true/false) - ignore chains composed entirely of residues other than the 20 standard amino acids
Ignore all non-standard residues (true/false) - ignore all residues other than the 20 standard amino acids

Ligand options - settings for the ligand preparation script (see limitations):

The ligand prep script will check for hydrogens and add them if they are missing. AutoDock Vina needs the polar (potentially H-bonding) hydrogens to identify atom types for scoring purposes.

Merge charges and remove non-polar hydrogens (true/false) - note AutoDock Vina does not use charges or nonpolar hydrogens, so this setting is not expected to affect results except for the presence or absence of nonpolar hydrogens in the ligand output files
Merge charges and remove lone pairs (true/false) - note AutoDock Vina does not use charges or lone pairs, so this setting is not expected to affect results except for the presence or absence of lone pairs in the ligand output files (and there may not have been any lone pairs to start with)

Advanced options - docking parameters

Number of binding modes (1-10, default 9) - maximum number of binding modes to generate
Exhaustiveness of search (1-8, default 8) - thoroughness of search, roughly proportional to time
Maximum energy difference (kcal/mol) (1-3, default 3) - maximum score range; binding modes with scores not within this range of the best score will be discarded

Executable location:

Opal web service (default)
- Server - URL of web service implemented using Opal; clicking Reset restores the URL for the service provided by the NBCR
Local
- Path - pathname of locally installed executable

OK initiates the calculation and dismisses the dialog, whereas Apply initiates the calculation without dismissing the dialog. The calculation is run as a background job. Clicking the information icon

in the Chimera status line will bring up the Task Panel, in which the job can be canceled if desired. Close dismisses the dialog, and Help opens this manual page in a browser window.

The docking results will be opened as a new model (with multiple submodels) and shown automatically using ViewDock. Please see the AutoDock Vina manual for a description of the output values.

Limitations

Docked peptides and other multiresidue ligands appear fragmented. AutoDock Vina changes the order of the atomic coordinates based on the branching of rotatable groups, and this scrambles multiresidue ligands such as peptides. The misordering of the atoms leads to an incorrect bonding pattern in Chimera. For all-atom representations (stick or ball-and-stick but not ribbons), there is a workaround: deleting all the bonds and adding them back based on distances. For example, if the docking results are opened as model #5, commands:

~ribbon #5
disp #5
~bond #5
sel #5

...and then adding all reasonable bonds among selected atoms using the Adjust Bonds section of Build Structure. Another workaround is to make the ligand a single residue before docking, but that would not allow display as a ribbon either.

Rings in ligand broken by misidentification of rotatable bonds. The ligand preparation script may identify bonds in rings as rotatable, resulting in impossible (“broken”) ring conformations. Because the workflow of the Chimera interface does not allow intervention between ligand preparation and docking, and there is no option to keep the ligand rigid, there is no workaround other than to prepare input files and run AutoDock Vina directly (not using Chimera). This ring-opening problem has been observed with prolines and may occur with other rings as well.

UCSF Computer Graphics Laboratory / August 2019