Re: [ccp4bb] Another puzzle: 5gnn: SOLVED

2016-09-09 Thread CHAVAS Leonard 
Could we write to them as a community? As the 'CCP4 Peer Reviewing Organisation 
for Structural Biology'?

Great job indeed!

Cheers, leo

-
Leonard Chavas
- 
Synchrotron SOLEIL
Proxima-I
L'Orme des Merisiers
Saint-Aubin - BP 48
91192 Gif-sur-Yvette Cedex
France
- 
Phone:  +33 169 359 746
Mobile: +33 644 321 614
E-mail: leonard.cha...@synchrotron-soleil.fr
-

> On 09 Sep 2016, at 13:39, Manfred S. Weiss 
>  wrote:
> 
> 
> Good job Randy!
> 
> Now, in order to finish this off properly somebody should
> refine the structure, alert the editor of the journal as well
> as the corresponding author of the paper.
> 
> If we leave it at that, the only thing we have is the satisfaction
> that we managed to "peer review" this structure, but nothing else is
> going to happen.
> 
> Cheers, Manfred
> 
> 
> On 09.09.2016 13:34, Gerard Bricogne wrote:
>> Dear Randy,
>> 
>> Congratulations for this terrific piece of work!
>> 
>> As James Fraser was writing two days ago, this case establishes a
>> new paradigm for the "peer review" of structural papers :-) .
>> 
>> 
>> With best wishes,
>> 
>>  Gerard.
>> 
>> --
>> On Fri, Sep 09, 2016 at 12:20:53PM +0100, Randy Read wrote:
>>> Dear all,
>>> 
>>> Following on from Paul’s suggestion that the space group is wrong, I 
>>> thought it might be interesting to try to solve this structure from 
>>> scratch, using what I’m currently teaching students at a CCP4 workshop.  
>>> What worked was to find some distant homologues with HHpred (top 3 hits 
>>> after ignoring 5gnn), make a trimmed ensemble, and search with that.  This 
>>> gave a very clear solution for one molecule in P65.  Model completion from 
>>> that point works really well, given the resolution, using either 
>>> phenix.autobuild or ARP/wARP.
>>> 
>>> For the edification of the members of the BB, the ARP/wARP model is 
>>> attached!  (Hopefully no flames for an attachment of 135kB…) This model 
>>> hasn’t had some really obvious rotamer fixes or extensions of the termini 
>>> applied, but it has good stereochemistry, and R/Rfree are 0.241/0.285.
>>> 
>>> Given the vastly improved statistics for a different space group and the 
>>> huge differences in the model, it is to be hoped that the authors retract 
>>> the original publication and PDB entry.  Thanks to Gerard for pointing out 
>>> the issues with this!
>>> 
>>> Best wishes,
>>> 
>>> Randy Read
>>> 
>> 
>> 
>>> 
>>> 
>>> -
>>> Randy J. Read
>>> Department of Haematology, University of Cambridge
>>> Cambridge Institute for Medical ResearchTel: +44 1223 336500
>>> Wellcome Trust/MRC Building Fax: +44 1223 336827
>>> Hills Road
>>> E-mail: rj...@cam.ac.uk
>>> Cambridge CB2 0XY, U.K.   
>>> www-structmed.cimr.cam.ac.uk
>>> 
 On 7 Sep 2016, at 16:51, Paul Adams  wrote:
 
 Dear Gerard,
 
 thanks for pointing this structure out. This is indeed very startling. The 
 paper indicates that the structure was refined with Phenix. However, if I 
 download the model/data and refine (with Phenix) the starting R-factors 
 are 0.36/0.38 and these get worse during refinement. Clearly the deposited 
 model is not consistent with the R-factors reported in the paper, or in 
 the wwPDB. In addition, analysis of the data suggests strongly that the 
 true symmetry is P6 (possibly with a screw axis) - I suspect that this is 
 the genesis of statements about twinning, but it isn’t clear if that was 
 used in the refinement. Being charitable I can imagine that the wrong 
 model was deposited in the wwPDB. However, this doesn’t explain the 
 R-factors reported by the wwPDB, or the less than convincing images of the 
 structure shown in the paper. I very much agree with you that there must 
 have been ample alarms sounded along the way. It is cautionary that this 
 wasn’t caught at some point. For me this highlights that the issues go 
 beyond the naivety or impatience of a single student.
 
 Cheers,
Paul
 
> On Sep 7, 2016, at 7:20 AM, Gerard Bricogne  
> wrote:
> 
> Dear all,
> 
>   While the thread on "Another MR pi(t)fall" is still lukewarm, and
> the discussion it triggered hopefully still present in readers' minds,
> I would like to bring another puzzling entry to the BB's attention.
> 
>   When reviewing on Monday the weekend's BUSTER runs on the last
> batch of PDB depositions, Andrew Sharff (here) noticed that entry 5gnn
> had been flagged as giving much larger R-values when re-refined with
> BUSTER (0.3590/0.3880) than the deposited ones (0.2210/0.2500). This
> led us to carry out some investigation of that entry.
> 
>   The deposited coordinates were flagged by BUSTER as having 4602
> bond-length violations,

Re: [ccp4bb] Another puzzle: 5gnn: SOLVED

2016-09-09 Thread Manfred S. Weiss 


Good job Randy!

Now, in order to finish this off properly somebody should
refine the structure, alert the editor of the journal as well
as the corresponding author of the paper.

If we leave it at that, the only thing we have is the satisfaction
that we managed to "peer review" this structure, but nothing else is
going to happen.

Cheers, Manfred


On 09.09.2016 13:34, Gerard Bricogne wrote:

Dear Randy,

 Congratulations for this terrific piece of work!

 As James Fraser was writing two days ago, this case establishes a
new paradigm for the "peer review" of structural papers :-) .


 With best wishes,

  Gerard.

--
On Fri, Sep 09, 2016 at 12:20:53PM +0100, Randy Read wrote:

Dear all,

Following on from Paul’s suggestion that the space group is wrong, I thought it 
might be interesting to try to solve this structure from scratch, using what 
I’m currently teaching students at a CCP4 workshop.  What worked was to find 
some distant homologues with HHpred (top 3 hits after ignoring 5gnn), make a 
trimmed ensemble, and search with that.  This gave a very clear solution for 
one molecule in P65.  Model completion from that point works really well, given 
the resolution, using either phenix.autobuild or ARP/wARP.

For the edification of the members of the BB, the ARP/wARP model is attached!  
(Hopefully no flames for an attachment of 135kB…) This model hasn’t had some 
really obvious rotamer fixes or extensions of the termini applied, but it has 
good stereochemistry, and R/Rfree are 0.241/0.285.

Given the vastly improved statistics for a different space group and the huge 
differences in the model, it is to be hoped that the authors retract the 
original publication and PDB entry.  Thanks to Gerard for pointing out the 
issues with this!

Best wishes,

Randy Read







-
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical ResearchTel: +44 1223 336500
Wellcome Trust/MRC Building Fax: +44 1223 336827
Hills RoadE-mail: 
rj...@cam.ac.uk
Cambridge CB2 0XY, U.K.   
www-structmed.cimr.cam.ac.uk


On 7 Sep 2016, at 16:51, Paul Adams  wrote:

Dear Gerard,

 thanks for pointing this structure out. This is indeed very startling. The 
paper indicates that the structure was refined with Phenix. However, if I 
download the model/data and refine (with Phenix) the starting R-factors are 
0.36/0.38 and these get worse during refinement. Clearly the deposited model is 
not consistent with the R-factors reported in the paper, or in the wwPDB. In 
addition, analysis of the data suggests strongly that the true symmetry is P6 
(possibly with a screw axis) - I suspect that this is the genesis of statements 
about twinning, but it isn’t clear if that was used in the refinement. Being 
charitable I can imagine that the wrong model was deposited in the wwPDB. 
However, this doesn’t explain the R-factors reported by the wwPDB, or the less 
than convincing images of the structure shown in the paper. I very much agree 
with you that there must have been ample alarms sounded along the way. It is 
cautionary that this wasn’t caught at some point. For me this highlights that 
the issues go beyond the naivety or impatience of a single student.

 Cheers,
Paul


On Sep 7, 2016, at 7:20 AM, Gerard Bricogne  wrote:

Dear all,

   While the thread on "Another MR pi(t)fall" is still lukewarm, and
the discussion it triggered hopefully still present in readers' minds,
I would like to bring another puzzling entry to the BB's attention.

   When reviewing on Monday the weekend's BUSTER runs on the last
batch of PDB depositions, Andrew Sharff (here) noticed that entry 5gnn
had been flagged as giving much larger R-values when re-refined with
BUSTER (0.3590/0.3880) than the deposited ones (0.2210/0.2500). This
led us to carry out some investigation of that entry.

   The deposited coordinates were flagged by BUSTER as having 4602
bond-length violations, the worst being 205.8 sigmas, and other wild
outliers. The initial Molprobity analysis gave a clash score of near
100, placing it in the 0-th percentile. The PDB validation report is
dominantly red and ochre, with only a few wisps of green.

   Examining the model and map with Coot showed "waters, waters
everywhere", disconnected density, and molecules separated by large
layers of water. The PDB header lists hundreds of water molecules in
REMARK 525 records that are further than 5.0 Angs from the nearest
chain, some of them up to 15 Angs away.

   The cartoons on the NCBI server at

http://www.ncbi.nlm.nih.gov/Structure/mmdb/mmdbsrv.cgi?uid=142582=1

show random coils threaded up and down through beta-strands, and the
one on the RCSB PDB site at

 http://www.rcsb.org/pdb/explore.do?structureId=5GNN

also shows mostly random coil, with only very few and very short