How Do I Prepare A Pdb For Submission To The Protein Data Bank?

Question

[Nick T]

I have a couple structures that are nearly ready to be deposited in the PDB. Out of curiosity, I ran them through the ADIT's precheck tool and they failed with one error after another, as I lacked all sorts of additional records (TER, SEQRES, HETNAM, etc.) that my refinement and modelling tools don't seem to care about.

From what I understand/guess, the online submission tool will help fill in all the metadata (REMARKs, etc.), but how do I turn my coordinates into something acceptable?

[Nick T]

(cross-post from biology.stackexchange.com)

Answer 1

[Egon Willighagen]

I am not entirely sure it is exactly what you are looking for, but WHATCHECK helps you create PDB files in that it can show irregularities, among which in coordinates.

Answer 2

[João Rodrigues]

Disclaimer: I've never submitted a structure to the PDB.

Have you tried pdb-extract?

Of the fields you mentioned SEQRES and HETNAM are user defined. Particularly SEQRES, it is impossible for a program to know what you had in the test tube. HIS tags, etc, these all have to belong there. HETNAM is kind of the same, being the chemical name of any ligand you use.. The PDB webpage mentions the following, and therefore I guess you have to write this information in yourself..

Gather information about the proteins/nucleic acids in your structure from sequence databases (e.g. using UniProtKB 1 or BLAST 2 ) or about your ligand (using Ligand Expo 3 ). Instructions will be provided during deposition to deposit new ligands. Collect log files from data processing, scaling, MR, refinement, etc. (pdb_extract can parse these files for you and gather sequence information for what was in the sample.)

TER statements on the other hand, should be handled by the modelling package you used. I'd follow Egon's advice and use WHATIF to generate a clear record of your coordinates. Those at least would be good.