By oversimplifying the situation here the PDB does not answer my
related point about competing crystallographers:
My scenario:
Group A deposits structure with new drug - gets their three-letter
code for example ZA3
they then get to check the coordinates and chemical definition of this ligand.
But suppose a little after that a competing group B deposits their
structure with the same drug which they think is novel - but no...
they get assigned the now described ZA3 which has been checked by the
other group.
Then it is a race to see who gets to publish and release first. And
if it is the second group B who wins then they are publishing the work
of their A competitors - who have done the depositing and checking of
the ligand description.
Sounds unlikely? Well, it actually happened in 2011 for my exact
example ZA3 - present in 2Y2I and in 2Y59 from competing groups.
From the dates in the mmcif it was 2Y2I depositors who set up and had
a chance to review the description of ZA3 ligand. Only to see it
released a week before their crystal structure, when their ZA3
appeared to accompany competing 2Y59! It is amazing that the PDB did
not spot this and arrange a suitable workaround.
Just to check:
mmcif for ZA3 shows it was created for 2Y2I:
...
_chem_comp.pdbx_model_coordinates_db_code 2Y2I
...
But it was modified for release:
...
_chem_comp.pdbx_modified_date 2011-07-22
...
corresponding to the early 2011-07-27 release date of the competing
structure: 2Y59 even though this PDB was _deposited_ second.
The ZA3 ligand definition released with 2Y59 actually embodies the
atomic coordinates from the 2Y2I structure:
<mmcif>
ZA3 O6 O6 O 0 1 N N N 8.279 7.165 40.963 0.311 -1.061 -0.920
O6 ZA3 1
ZA3 C5 C5 C 0 1 N N N 9.132 8.047 40.908 0.147 -0.205 -0.073
C5 ZA3 2 ...
<PDB 2Y2I>
HETATM 3598 O6 ZA3 A1000 8.279 7.165 40.963 1.00 41.25 O
HETATM 3599 C5 ZA3 A1000 9.132 8.047 40.908 1.00 63.20
C ...
Surely a better approach would be to allow both groups a chance to
work through and sign off on independent ligand descriptions?
Then whoever releases first would release both a novel structure and
the ligand definition _they_ deposited and checked. Their priority can
then be asserted and the other group contacted to ask if they agree to
accept this definition. This also has the advantage of better
confidentiality pre-publication.
Another problem from any cross-linking of definitions is that say
group A are motivated by reviewers' reports to change the definition
of ZA3 pre-release. Well now the change impinges on the chemical
meaning of other group B's deposited structure. For example ZA3 mmcif
has a statement:
ZA3 "Modify aromatic_flag" 2011-06-04 RCSB
so this change was pre-release - but we cannot be sure what motivated
this - whether it was signed off by the 2Y2I authors or the 2Y59
authors (or both?)....
With the accelerating pace of drug discovery for sure this sort of
uncertainty is going to happen again.Unless the PDB have changed their
practice for ligand deposition?
All the best
Martyn
Cambridge.
On Fri, Jun 19, 2015 at 1:49 PM, Sheriff, Steven <[log in to unmask]> wrote:
> All:
>
>
>
> Since the original query was cross-posted on both the COOT mailing list and
> the CCP4BB Rachel Green gave me permission to forward this to both. She
> provides links about the mechanism of assignment of 3-letter codes. In the
> third link below, my original suggestion to the COOT mailing list that one
> could just use UNK is incorrect as that is reserved for unknown amino acids.
> According to this document, I should have suggested UNL for an unknown
> ligand.
>
>
>
> Steven
>
>
>
> From: Rachel Kramer Green [mailto:[log in to unmask]]
> Sent: Tuesday, June 16, 2015 10:21 AM
> To: Sheriff, Steven
> Cc: info
> Subject: Re: New ligand 3-letter code (help-7071)
>
>
>
> Dear Steven,
>
> During annotation of ligands, all chemical components present in the
> structure are compared against the definitions in the Chemical Component
> Dictionary (http://www.wwpdb.org/data/ccd). If the ligand is not in the
> dictionary, a three letter code is assigned. See
> http://www.wwpdb.org/documentation/policy#toc_assignment. In the future, a
> group of three-letter codes may be set aside to be used during refinement to
> flag new ligands.
>
> Clarification about the ligand ids assignment and in particular the usage of
> UNX/UNL/UNK residues can be found at
> http://www.wwpdb.org/documentation/procedure#toc_2.
>
> Best wishes,
> Rachel
>
>
>
> ________________________________
>
> Rachel Kramer Green, Ph.D.
>
> RCSB PDB
>
> [log in to unmask]
>
>
>
> New! Deposit X-ray data with the wwPDB at:
>
> http://deposit.wwpdb.org/deposition (NMR and 3DEM coming soon).
>
> ___________________________________________________________
>
> Twitter: https://twitter.com/#!/buildmodels
>
> Facebook: http://www.facebook.com/RCSBPDB
>
>
>
>
>
>
>
> On 6/5/2015 7:50 AM, Sheriff, Steven wrote:
>
> All:
>
>
>
> Why the concern for unassigned three-letter codes? The wwPDB isn’t going to
> let you assign a three-letter code, it will choose its own code.
>
>
>
> At BMS (a pharmaceutical company), we do many hundreds of structures a year
> with ligands and we assign the same, already assigned, three-letter code for
> all of our ligands (unless we have two or more different ligands in a single
> structure, in which case we use two or more different already assigned
> three-letter codes). COOT can mostly handle this.
>
>
>
> However, I believe that if you want an unassigned code, the wwPDB has set
> aside UNK[nown] for this purpose.
>
>
>
> Steven
>
>
>
> From: Mailing list for users of COOT Crystallographic Software
> [mailto:[log in to unmask]] On Behalf Of Eleanor Dodson
> Sent: Friday, June 05, 2015 6:28 AM
> To: [log in to unmask]
> Subject: Re: New ligand 3-letter code
>
>
>
> I use your method - trial & error..
>
> It would be nice if at least there was a list somewhere of unassigned codes!
>
>
>
> On 5 June 2015 at 09:16, Lau Sze Yi (SIgN)
> <[log in to unmask]> wrote:
>
> Hi,
>
>
>
> What is the proper way of generating 3-letter code for a new ligand? As of
> now, I insert my ligand in Coot using smiles string and for the 3-letter
> code I picked a non-existent code by trial and error (not very efficient). A
> cif file with corresponding name which I generated using Phenix was imported
> into Coot.
>
>
>
> I am sure there is a proper way of doing this. Appreciate your feedback.
>
>
>
> Regards,
>
> Sze Yi
>
>
>
> ________________________________
>
> This message (including any attachments) may contain confidential,
> proprietary, privileged and/or private information. The information is
> intended to be for the use of the individual or entity designated above. If
> you are not the intended recipient of this message, please notify the sender
> immediately, and delete the message and any attachments. Any disclosure,
> reproduction, distribution or other use of this message or any attachments
> by an individual or entity other than the intended recipient is prohibited.
>
>
>
> ________________________________
> This message (including any attachments) may contain confidential,
> proprietary, privileged and/or private information. The information is
> intended to be for the use of the individual or entity designated above. If
> you are not the intended recipient of this message, please notify the sender
> immediately, and delete the message and any attachments. Any disclosure,
> reproduction, distribution or other use of this message or any attachments
> by an individual or entity other than the intended recipient is prohibited.
