I had thought that after discussion on this bulletin board we could find no 
rational scientific reason why the PDB renumbers waters from the N to the C 
term of the polypeptides (and 5' to 3' of nucleic chains?). 
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1311&L=CCP4BB&D=0&P=12314

I think that provided you follow the agreed rules on chain ids (see below) then 
you have a 'right of veto' on the water renumbering. You also have a 'right of 
veto' for the water movement to symmetry-related positions - which would 
involve chain id change. 

So probably the best way is to assign the chain ids yourself and just say you 
don't want any of your waters renumbered. Then keep track of them yourself as 
Ed suggested using documented community tools.

One related question is how waters with alternate positions are dealt with. I 
cannot find clear documentation on this. Often sidechains with alternate 
positions motivate alternate waters during refinement. But I believe these are 
removed during the deposition water renumbering step. So if you have particular 
waters in alternative positions - say owing to a Michaelis complex at the 
active site - then the renumbering will likely split your water in alternate 
positions to produce two apparently unrelated, differently-numbered, solvent 
atoms. It could be argued how can you know these are alt pos of the same water? 
But if you have refined them like that then it is good for users of your 
coordinates to see that documented.  

It may be that the new PDB Deposition System is more flexible in this regard - 
I have not used it yet.

Below is the current agreed procedure from the wwPDB website.

Hope that helps. 
 regards
  Martyn 
How are chain IDs assigned to chemical groups and waters?
All chemical groups and waters within 5 Ångstroms of any atom of a polymeric 
chain will be considered to be associated with that chain and will be assigned 
the same chain ID as that polymer. If a particular chemical group/water is 
equidistant from more than one chain, then the chain ID is randomly assigned to 
be that of any one of these polymers.
Waters further than 5 Ångstroms away from the polymer that can be moved by 
symmetry to within 5 Ångstroms, will then be automatically moved and the author 
will be notified. If there are objections, the waters will be moved back to 
their original positions.
Waters further than 5 Ångstroms away from any polymer, which cannot be brought 
closer to a polymer chain by application of symmetry, will be brought to the 
attention of the depositor. The waters will be listed in REMARK 525 with the 
distance to the nearest macromolecule listed. Authors will be given the 
opportunity to remove the waters or update the coordinates. Authors may choose 
to retain the distant waters in the entry.


________________________________
 From: Edward A. Berry <[email protected]>
To: [email protected] 
Sent: Monday, 7 April 2014, 15:25
Subject: Re: [ccp4bb] Renumbering of water molecules during deposition
 

After deposition (my experience is with RCSB) you will get back a processed
file to verify.  See how the waters have been renumbered, and renumber
in such a a way that your original scheme is easily identifiable but in
keeping with whatever the renumbering was trying to achieve- e.g. your
waters X1,X2,X3 could become A1001, A1002, A1003 and the symmetry related
waters B1001, B1002, B1003. Discuss with the annotator to see how your
waters can be renumbered in a recognizable way in keeping with requirements
of the PDB conventions. Then in future depositions keep this same scheme
(you may still need to renumber the original processed file each time).
eab

Eckhard Hofmann wrote:
> Hi all,
> during deposition of structures the water numbering is frequently changed.
> I would love to keep specific numbers for critical water molecules e.g.
> in an active site in related proteins (mutants, ligand soaks), to
> facilitate easier discussion in a manuscript.
>
> Is there a general agreement how to treat these?
>
> One way to deal with this is would be to use an arbitrary offset:
> A1-A233 protein
> A301 ligand
> A302 another ligand
> A401-404 4 special waters
> A500- general solvent
>
> Any input welcome ;-)
> All the best,
> Eckhard
> --
> Prof. Dr. Eckhard Hofmann <[email protected]>
> Ruhr-Uni Bochum
> AG Proteinkristallographie, LS Biophysik, ND04/318
> 44780 Bochum
> Tel: +49-(0)234/32-24463, Sekr. -24461, FAX: -14762
>

Reply via email to