Thursday, 10 June 2010

How to get an experimental ligand structure from the PDB?

The Blue Obelisk Q&A website at is really turning into a useful resource. There are now over 100 user accounts, and the quality of the questions and answers is generally quite high. Although several of the questions relate to open source cheminformatics software, many are simply general cheminformatics questions. You can log in with OpenID or your Google account, or even ask questions anonymously (but this doesn't work so well, as you won't be updated when it's answered).

Q&A sites are much better than forums or mailing lists for handling technical questions because there is no repetition; each question only gets asked once (how times have you seen the same question on the CCL for example?). If additional information becomes available, that can be added as a new answer (unlike a forum, where adding to an old forum thread is discouraged). Also, if a question is ambiguous, the original author can correct it after some prompting (in a forum, the corrected question may appear half-way down the page).

Anyway, I recently asked the following question which is related to my research:
Given a PDB code, what is the simplest way to get an SDF file with the experimental coordinates of a particular ligand?

My current procedure is to go to the PDB Ligand Expo, and use the “Search for instances of chemical components by 3-letter ID code” at the bottom with “PDB entry codes + coordinates files” chosen. This requires me to specify the 3-letter ID code, which I need to have obtained from the web page for the PDB file.

On the search results page, I can simply click on the SDF file for the particular PDB file in which I am interested to download the experimental coordinates.

Is there an easier way? For example, it does not seem to be possible to go directly from the website for a particular PDB entry.

(I would also be interested in the same question applied to a handling a large number of PDB codes at the same time.)
Matteo Floris provided the answer:
First, you should download this dictionary (Tabulation of PDB entries containing each chemical component); in this way you can get the full list of ligands associated to PDB entries.
Then, the best is to download the SDF file with all the ligands and use the dictionary for the mapping PDB identifier -> list of associated ligands.

...this gives unique combinations of PDB entry – ligand – PDB entry chain… let me show two examples: the first is the ligand ZU3, for which we have the coordinates in all the PDB where this ligand is present:

2zu3_ZU3_1_A_501B 2zu4_ZU3_1_A_501B

the second example, the ligand ZZC, that is present in the SDF file for the same PDB entry but in 2 different chains (A and B):

2wod_ZZC_1_A_401D 2wod_ZZC_1_B_401F

Now to test it out!

1 comment:

Egon Willighagen said...

There are about 40 unique visitors each day.