Getting Started

Revision as of 06:37, 13 March 2014 by Frodo (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Getting Started with Docking

There are exceptions to every rule. Please do not use this guide as a substitute for common sense.

Docking requires a structure of the target to dock to.

Target structure NOT available

You may still be able to dock to a homology model of your target, if you know the sequence. Try using ModBase. If that fails, you may still be able to use complementary methods such as shape similarity, 2D scaffold hopping. Failing that, you can still fall back on experimental methods, including high throughput screening and crystallographic fragment screening.

Target structure IS available

If you have only one structure available, then that is probably the one you should use. If you have more than one structure, then you will probably need to pick a single structure to dock to. You may also be able to use multiple structures in various ways to inform your docking.

Generally, ligand-bound crystal structures are preferable to apo-enzyme crystal structures, which in turn are preferable to NMR structures, which in turn are preferable to homology or hand-built models. High resolution, problem free structures are better than lower resolution structures, or structures with problems.

Before docking, you should take a close look at the features of your target structure.

You should collect together any activity data you may have.

You should consider which database subset to dock.

You should perform controls.

You should search for similar compounds, both commercially available and in annotated databases.

You should perform a small pilot docking study to test the calibration of your docking model.

You should review the output from docking, and fine tune the docking model if appropriate.

Recipes & Checklists

Have one ligand bound crystal structure only

Have one apo-enzyme form structure only

Have a good homology model

Only have a shaky homology model