DOCK Blaster:Prepare Input

From DISI
Revision as of 12:02, 8 October 2012 by Therese (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This page describes how to prepare input for and then to run DOCK_Blaster, a free web-based virtual screening service. We assume here that you have dealt with the important preliminary considerations and have a target structure and some indication of the binding site in hand. You optionally have additional information about actives and inactives for this target. This page provides a quick overview of preparing input files for docking. It also references more detailed information in separate articles. To follow the progress of a job after submitting it, please see DOCK Blaster:Job Management. Sample data are available if you do not wish to prepare your own files.

Quick Overview

The receptor must be in PDB or mol2 format. A docked ligand in mol2 format (e.g. a crystallographically observed or modeled ligand) or a binding site specification in PDB format, which may be "hot spots" or atoms of binding site residues is also required. You may optionally specify up to 50 actives and up to 50 inactives as SMILES or ZINC IDs. Larger databases (e.g. lead-like, fragment-like) may only be docked after initial control calculations have completed successfully.

More detailed instructions

  • 1. Specify a target in PDB or mol2 format.

The PDB file must have the following characteristics, to avoid confusing any of the programs used by DOCK Blaster. We recommend PDB, but do also support mol2.

    • a. Only ATOM cards are allowed (HETATM are mapped to ATOM)
    • b. Chain identifier should be blank
    • c. If there is a cofactor, you should specify a dictionary of parameters for it. A number of co-factor dictionaries are available online. If you do not supply a cofactor dictionary, it will be treated as occupying space and be electrostatically inert. Also, you will receive warning messages that this treatment has been performed.
    • d. Residue numbers must monotonically increase. This is particularly important if there are multiple domains, for example, with A and B chain identifiers. Our interface automatically renumbers residues if required, and warns you that this has been done.
    • e. Disorder models must be resolved into a single choice for each atom. If you do not resolve them, our script will automatically take the 'A' choice, and warn you that this has been done.
    • f. If you are docking to a metalloenzyme, special steps are required.
    • g. Waters must be removed or made explicit. Any structural waters should be "TIP" as "HOH" waters will be removed automatically.
    • h. There are numerous additional steps taken to process the receptor, which we will write here as we think of them / add them.

If you do not have your own target, you may use some of our sample data or try one of our tutorials.

If you upload the target in mol2 format, it will be converted to PDB. Certain additional conventions in the mol2 file must be followed for this to work correctly. Please see the test data for examples.

  • 2. Specify the binding site.

You may specify the binding site using a docked ligand in mol2 format, or residues bounding the binding site in PDB format. If you upload a mol2 file, it will be scored, redocked, and used in various other ways to evaluate the quality of the docking model. If you do not have a ligand (experimental or modeled) then you must specify at least 3 residues, the center of mass of which is at the center of the binding site.

  • 3. Specify your email address if you wish to receive status updates by email. To protect the anonymity of our users, DOCK Blaster does not require registration. Furthermore, we will not give out any personal information we may acquire about you. The email address here is simply there to be helpful to you, but it is optional.
  • 4. If you have actives and inactives, you may specify them in the fields provided. Currently we support uploading up to 50 of each. They will be docked and scored, and can thus offer immediate feedback on the performance of the docking calculation.
  • 5. If you wish to keep your calculation private, specify a PIN. The data you upload and produce will only be accessible with this PIN. If you wish to make your results available to anyone, leave the PIN blank. Note that we cannot guarantee the confidentiality of any results.
  • 6. We recommend you write notes about the calculation in the memo field provided. This is particularly helpful if you launch many calculations. We recommend you keep a notebook, and note each Job ID down to keep track of your calculations.
  • 7. If your target has a cofactor (other than a metal), you will need to specify a cofactor. Many cofactors are available from our co-factor dictionary. The cofactor must match all the atoms in your target structure or the calculation will not start.
  • 8. Experts have the option to provide additional information to the docking calculation. No public documentation is available about this at this time. If you encounter problems with docking, we may provide an expert.tar file to help you work around weaknesses in the current design.
  • 9. If you are happy with the input and the terms and conditions, you may click on "I agree. DOCK!" to submit the calculation. What happens next is describe in DOCK Blaster:Job Management.


Yet More Detailed Information

and the DOCK Blaster main page.