ADapt

The AutoDock preparation & perception tool: from a PDB entry to receptor and ligand files ready for docking with the AutoDock docking engines, in a web browser.

ADapt with PDB entry 1FV1 loaded: the residue tree on the left, the 3D view on the right
ADapt with PDB entry 1FV1. The asymmetric unit (ASU) tab is open; the numbered parts are explained below.

The problem

Docking programs have become fast and accessible enough to be used by a broad community, including occasional users. Their results, however, depend on the quality of the input. The receptor must be given to AutoDock in its own PDBQT format, and building that file from the data the Protein Data Bank provides is neither obvious nor free of pitfalls. A PDB file records a crystallographic experiment, not a molecule ready for docking. To prepare a receptor from it, you must understand what the file contains and make a series of choices:

Most of these choices depend on where a fragment sits relative to the pocket you intend to dock into, so they are best made while looking at the structure.

What ADapt does

ADapt reads a PDB file, or fetches it from the RCSB, and analyses it with ADFR's own code, the same code as the PDB analyzer of AutoGridFR's graphical interface (AGFRgui):

What is known to be wrong or missing, in ADapt and in the ADFR code it runs, is listed on one page: Known issues.

The receptor files ADapt writes are byte-identical to those of ADFR's command line (agfr --toPdbqt) for the same choices, except for one line ADapt adds to each file to record which biomolecule was used. ADapt's test suite checks this.

The interface

  1. Open PDB… reads a local .pdb, .ent, .pdb.gz or .ent.gz file. Dropping a file on the page does the same.
  2. Fetch from RCSB: type a PDB id and press Enter.
  3. The entry, its experiment and its resolution.
  4. One tab per construct: the ASU and every biological assembly (Biomol 1, 2, …). The tab selects the molecule that will be written.
  5. The tree: chains, then categories of residues, then residues. The first checkbox column puts a fragment in the receptor.
  6. params: per-residue settings (alternate location, rotamer, ion charge).
  7. lig.: the second checkbox column puts a fragment in the ligand.
  8. show rec. / lig. / other highlights, in the 3D view, what is currently in the receptor, in the ligand, or in neither.
  9. receptor PDBQT and ligand PDBQT write the files and download them.
  10. The 3D view (3Dmol.js). Receptor atoms have grey carbons, ligand atoms orange sticks; symmetry copies are drawn translucent; missing residues are drawn in orange. Clicking a tree node highlights its atoms in green and centres the view on them (see The tree and the 3D view).
  11. The status line: what was loaded, highlighted or written.
  12. Files written (with download links and the session folder) and the Log of ADFR's messages.

Right-clicking a residue opens its menu: complete a side chain or leave it as is, or mutate the residue. The Docs button at the top right opens this documentation.

Fragment categories

Under each chain, residues are grouped by category. A category's checkbox shows whether its residues are in the receptor: checked, unchecked, or partly checked. Clicking it applies to all of them.

CategoryWhat it holdsIn the receptor by default
proteinComplete standard amino acids.yes
nucleicNucleic acid residues.yes
sidechains to completeStandard amino acids whose missing side-chain atoms can be rebuilt from the rotamer library.yes, with the best-scoring rotamer
missing atomsOther residues with missing atoms (for instance backbone atoms).yes, as they are
cofactorsResidues in ADFR's table of cofactors.yes
modifiedResidues the file declares as modified (MODRES records).yes
manual chargesSingle-atom ions that AutoDock supports but that the Gasteiger charge model cannot charge (Zn, Mg, Ca, Fe, Mn; also Br). They get a formal charge (see metals).no
ligandsResidues that ADFR's table of PDB entries lists as ligands of this entry.no
additivesCrystallization additives: GOL, EDO, SO4, PO4, PEG, DMS, ACT and others.no
waterWater molecules.no
otherAnything not in the categories above.no

A chain can also have three sub-groups:

Cases

Each page below walks through one situation with a PDB entry that shows it. Every entry can be fetched by its id to follow along.

Biological assembly, not the ASU1HHP, HIV-1 protease: the ASU holds half of the enzyme. Choosing between assemblies1FV1: two copies of one complex in one file. Missing side chainsRebuild them from rotamers, or leave them as they are. Alternate locationsPick the conformation to dock against. Ligands and other hetero groups6GGD: take the ligand out of the receptor. Metals, ions, unsupported atomsA zinc charge; a copper AutoDock cannot type. Missing residues4A1X: gaps and missing termini, in the tree and in 3D. Mutating residuesReplace a residue by any standard amino acid. Writing the PDBQT filesWhat is written, and how it is recorded. Provenance in the PDBQT filesEvery REMARK record, and how to rebuild a receptor from it.

References

Getting started →