Re: [ccp4bb] Pseudo translational symmetry Problem or Wrong Spacegroup

[Edward A. Berry]

Does this mean that, when placing the first (or only) copy of a model,
the score will not be elevated by presence of tNCS (Unless the model contains
domains separated by more or less exactly the distance of the tNCS vector)?
(In general, using a generic MR program that does not correct for TNS.)
So a good score at the first step is likely to be correct, even if there is 
tNCS?
eab


Randy Read wrote:

Dear Yurong,

I just wanted to check that you're using an up-to-date version of Phaser, which 
will account for the presence of
translational NCS (tNCS). With older versions of Phaser, you could get apparent 
solutions that were incorrect but would
give a high LLG just because they satisfied the tNCS. However, even with the 
current version of Phaser there's a
potential problem here. Phaser will only take account of the tNCS if you're 
looking for an even number of copies (so
that it can look for pairs related by the tNCS). I'm wondering if, in your 
case, the tNCS is generated in the model by
placing the heterotetramer with its internal 2-fold parallel to the 2(1) screw 
axis along y. If that's the case, and
you're only looking for one copy of the heterotetramer, then I'm afraid Phaser 
will ignore the tNCS. If indeed you're
only looking for one copy of the heterotetramer, could you try looking for two 
copies of the half-tetramer instead (in
the new version of Phaser)? That may give a clearer indication of the true 
symmetry and, if the solution is correct, the
combination of translational symmetry and the screw axis should generate a full 
tetramer.

In cases like this, there's always a potential issue with twinning. I'd be 
interested in seeing the logfile (off-list)
for the Phaser run, which should give an indication of whether the intensity 
moments suggest twinning, once the
statistical effect of tNCS has been accounted for.

Best wishes,

Randy Read

On 3 Dec 2012, at 11:49, Yurong Wen wrote:


Dear All,

Recently, I collected a dataset of a protein-DNA complex and indexed in 
spacegroup P212121 to 3.4 Å. However,
Phenix.Xtriage indicated the presence of a very high pseudotranslational 
symmetry peak. I scaled the data in
spacegroup P222. When I used the protein heterotetramer as search model to do 
the MR, solutions are suggested in
P22121, P212121, P2212, P21212 and all with a TFZ score higher than 15, LLG 
higher than 570. Then I used phenix.refine
to refine those solutions; however the Rfree is as high as 0.54-0.56. The 
refinement strategy that I used for the
refinement is rigid body, group B-factors and XYZ coordinates.

How to deal with this pseudotranslational symmetry problem? Does a solution 
with such high TFZ scores mean that the
correct one has been found? How to solve the spacegroup problem in such 
situation?

Do you have any suggestions?

Thank you very much.
Greetings,
Yurong


--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research Tel: + 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

Re: [ccp4bb] Pseudo translational symmetry Problem or Wrong Spacegroup

[Randy Read]

Dear Yurong,

I just wanted to check that you're using an up-to-date version of Phaser, which 
will account for the presence of translational NCS (tNCS).  With older versions 
of Phaser, you could get apparent solutions that were incorrect but would give 
a high LLG just because they satisfied the tNCS.  However, even with the 
current version of Phaser there's a potential problem here.  Phaser will only 
take account of the tNCS if you're looking for an even number of copies (so 
that it can look for pairs related by the tNCS).  I'm wondering if, in your 
case, the tNCS is generated in the model by placing the heterotetramer with its 
internal 2-fold parallel to the 2(1) screw axis along y.  If that's the case, 
and you're only looking for one copy of the heterotetramer, then I'm afraid 
Phaser will ignore the tNCS.  If indeed you're only looking for one copy of the 
heterotetramer, could you try looking for two copies of the half-tetramer 
instead (in the new version of Phaser)?  That may give a clearer indication of 
the true symmetry and, if the solution is correct, the combination of 
translational symmetry and the screw axis should generate a full tetramer.

In cases like this, there's always a potential issue with twinning.  I'd be 
interested in seeing the logfile (off-list) for the Phaser run, which should 
give an indication of whether the intensity moments suggest twinning, once the 
statistical effect of tNCS has been accounted for.

Best wishes,

Randy Read

On 3 Dec 2012, at 11:49, Yurong Wen wrote:

> Dear All,
> Recently, I collected a dataset of a protein-DNA complex and indexed in 
> spacegroup  P212121 to 3.4 Å. However,  Phenix.Xtriage indicated the presence 
> of a  very high pseudotranslational symmetry peak.  I scaled the data in 
> spacegroup P222. When I used the protein heterotetramer as search model to do 
> the MR, solutions are suggested in P22121, P212121, P2212, P21212 and all 
> with a TFZ score higher than 15, LLG higher than 570. Then I used 
> phenix.refine to refine those solutions; however the Rfree is as high as 
> 0.54-0.56. The refinement strategy that I used for the refinement is rigid 
> body, group B-factors and XYZ coordinates.
> 
> How to deal with this pseudotranslational symmetry problem? Does a solution 
> with such high TFZ scores mean that the correct one has been found? How to 
> solve the spacegroup problem in such situation?
> 
> Do you have any suggestions?
> 
> Thank you very much.
> Greetings,
> Yurong
>  

--
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research  Tel: + 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

Re: [ccp4bb] Pseudo translational symmetry Problem or Wrong Spacegroup

[Herman . Schreuder]

Dear Yurong,
 
your approach to individually refine each potential solution is correct. I do 
not know how things are implemented in phenix.refine, but the first thing to 
check is that your are using the correct space group as obtained from the MR 
program. Some refinement programs use the space group from the reflection file 
(MTZ), which may be the space group used for the processing instead of the 
space group from the pdb file found by the MR program. 
 
>From what you write, you must have non-crystallographic translations of almost 
>0.5 in x and z direction. This means that you will have along these axis 
>either exact 21 symmetry, where the ~0.5 shift will generate a pseudo 2-fold, 
>or you have exact 2-fold symmetry where the pseudo symmetry will produce a 
>pseudo 21 axis. With your low resolution data, this explains why you get 
>solutions in all these space groups. You really have to try all combinations 
>of crystallographic and non-crystallographic symmetry.
 
Good luck!
Herman 
 




From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of 
Yurong Wen
Sent: Monday, December 03, 2012 12:49 PM
To: CCP4BB@JISCMAIL.AC.UK
Subject: [ccp4bb] Pseudo translational symmetry Problem or Wrong 
Spacegroup


Dear All,

Recently, I collected a dataset of a protein-DNA complex and indexed in 
spacegroup  P212121 to 3.4 Å. However,  Phenix.Xtriage indicated the presence 
of a  very high pseudotranslational symmetry peak.  I scaled the data in 
spacegroup P222. When I used the protein heterotetramer as search model to do 
the MR, solutions are suggested in P22121, P212121, P2212, P21212 and all with 
a TFZ score higher than 15, LLG higher than 570. Then I used phenix.refine to 
refine those solutions; however the Rfree is as high as 0.54-0.56. The 
refinement strategy that I used for the refinement is rigid body, group 
B-factors and XYZ coordinates.

How to deal with this pseudotranslational symmetry problem? Does a 
solution with such high TFZ scores mean that the correct one has been found? 
How to solve the spacegroup problem in such situation?

Do you have any suggestions?

Thank you very much.
Greetings,
Yurong
 

Re: [ccp4bb] pseudo translational symmetry

[Eleanor Dodson]

Brilliant illustration - even in 3D!
 eleanor
On 3 Aug 2012, at 14:01, Laurie Betts wrote:

> I couldn't resist this shot. It might not make the cut.but
> 
> 

Re: [ccp4bb] pseudo translational symmetry

[D Bonsor]

"Some people might squawk at 2MB attachments on a BB..."

Surely that is bark :)

Re: [ccp4bb] pseudo translational symmetry

[David Schuller]

One way to address this would be to provide a link to the image on the 
web, rather than including it as an attachment.



On 08/03/12 14:04, Francis E Reyes wrote:

Some people might squawk at 2MB attachments on a BB...



But that made my day..

Thanks!

On Aug 3, 2012, at 6:01 AM, Laurie Betts wrote:


I couldn't resist this shot. It might not make the cut.but





--
===
All Things Serve the Beam
===
   David J. Schuller
   modern man in a post-modern world
   MacCHESS, Cornell University
   schul...@cornell.edu

Re: [ccp4bb] pseudo translational symmetry

[Francis E Reyes]

Some people might squawk at 2MB attachments on a BB...



But that made my day..

Thanks!

On Aug 3, 2012, at 6:01 AM, Laurie Betts wrote:

> I couldn't resist this shot. It might not make the cut.but
> 
> 

Re: [ccp4bb] pseudo translational symmetry

[Jacob Keller]

What symmetry operator is that towel indicating? Must be 24-fold rotational
NCS?

JPK

On Fri, Aug 3, 2012 at 8:43 AM, Bosch, Juergen  wrote:

> Thank Laurie,
>
> what's that domain dangling off your DOG-domain ? And have you checked if
> the off-origin peak is indeed the same height as the origin peak ?
>
> May I include your slide in my class ?
>
> Jürgen
>
>
> On Aug 3, 2012, at 9:01 AM, Laurie Betts wrote:
>
> I couldn't resist this shot. It might not make the cut.but
> 
>
>
> ..
> Jürgen Bosch
> Johns Hopkins University
> Bloomberg School of Public Health
> Department of Biochemistry & Molecular Biology
> Johns Hopkins Malaria Research Institute
> 615 North Wolfe Street, W8708
> Baltimore, MD 21205
> Office: +1-410-614-4742
> Lab:  +1-410-614-4894
> Fax:  +1-410-955-2926
> http://lupo.jhsph.edu
>
>
>
>
>


-- 
***
Jacob Pearson Keller
Northwestern University
Medical Scientist Training Program
email: j-kell...@northwestern.edu
***

Re: [ccp4bb] pseudo translational symmetry

[Bosch, Juergen]

Thank Laurie,

what's that domain dangling off your DOG-domain ? And have you checked if the 
off-origin peak is indeed the same height as the origin peak ?

May I include your slide in my class ?

Jürgen


On Aug 3, 2012, at 9:01 AM, Laurie Betts wrote:


I couldn't resist this shot. It might not make the cut.but



..
Jürgen Bosch
Johns Hopkins University
Bloomberg School of Public Health
Department of Biochemistry & Molecular Biology
Johns Hopkins Malaria Research Institute
615 North Wolfe Street, W8708
Baltimore, MD 21205
Office: +1-410-614-4742
Lab:  +1-410-614-4894
Fax:  +1-410-955-2926
http://lupo.jhsph.edu

Re: [ccp4bb] pseudo-translational symmetry

[Eleanor Dodson]

Oops - sorry to mislead you.
Yes - you are absolutely right..
Eleanor


ale...@pasteur.fr wrote:

Just a small comment on Eleanor's instructive advices :

  

...
If your original structure has cell

(a2= 47.3, b2=58.9, c2=67.6) and angles (alpha2=90, beta2=99.2, gamma2=90)

 and the second cell is
(a1=67.5, b1=58.8, c1=98.9) and angles (alpha1=90, beta1=101.5, gamma1=90)
...
I would force my second cell to be
 98.9  58.8   67.5 (alpha1=90, beta1=101.5, gamma1=90) by reindexing l h -k



I guess the proper reindexing would be l k -h, wouldn't it?
(l h -k would not really do what you want AND it would also change the
reference system to a left-handed one...)

Cheers,
Alejandro




--
Unit of Protein Crystallography
Institut Pasteur de Montevideo
Uruguay


  

Re: [ccp4bb] pseudo-translational symmetry

[alebus]

Just a small comment on Eleanor's instructive advices :

> ...
> If your original structure has cell
>
> (a2= 47.3, b2=58.9, c2=67.6) and angles (alpha2=90, beta2=99.2, gamma2=90)
>
>  and the second cell is
> (a1=67.5, b1=58.8, c1=98.9) and angles (alpha1=90, beta1=101.5, gamma1=90)
> ...
> I would force my second cell to be
>  98.9  58.8   67.5 (alpha1=90, beta1=101.5, gamma1=90) by reindexing l h -k

I guess the proper reindexing would be l k -h, wouldn't it?
(l h -k would not really do what you want AND it would also change the
reference system to a left-handed one...)

Cheers,
Alejandro




--
Unit of Protein Crystallography
Institut Pasteur de Montevideo
Uruguay

Re: [ccp4bb] pseudo-translational symmetry

[Eleanor Dodson]

First - there doesnt seem much to worry about. The Rs will be higher 
than usual when you have a strong pseudo-translation vector. There are 
many weak observations for the h k l=2n+1 reflections.


But in cases like this is is very helpful to force the same indexing on 
all your different data sets so you can compare results easily:


If your original structure has cell

(a2= 47.3, b2=58.9, c2=67.6) and angles (alpha2=90, beta2=99.2, gamma2=90)

and the second cell is 
(a1=67.5, b1=58.8, c1=98.9) and angles (alpha1=90, beta1=101.5, gamma1=90)


with a pseudo translation of 0 0 0.5, then they are pretty obviously related.
I would force my second cell to be 
98.9  58.8   67.5 (alpha1=90, beta1=101.5, gamma1=90) by reindexing l h -k 
then you can just use the coordinate solution from the first cell plus a translated molecule.


Ditto when processing in P1 make sure the cell is 
98.9  58.8   67.5 89.9 101.5 89.9 and generate the 8 molecules 

Twinning checks can be misled when you have assigned the lower symmetry spacegroup - ie sent it P1 data instead of P21. 
And similarly they are less reliable when you have a strong pseudo-translation vector. 


Eleanor





Hi all,

I am trying to refine a structure to about 2.0A.  Indexing in HKL2000 indicates 
the protein crystallized in P2 with unit cell lengths (a1=67.5, b1=58.8, 
c1=98.9) and angles (alpha1=90, beta1=101.5, gamma1=90).  Molecular replacement 
with Phaser yields a solution in P1 21 1 with four molecules in the AU.  
Successive rounds of refinement lead to good quality maps with chemically 
appropriate packing of symmetry-related molecules.  However, TLS and restrained 
refinement lead to continuing drops of Rfactor (around 20%, depending on 
weighting) but the stalling of Rfree around 27%.

When the dataset is processed in P1, the unit cell has the dimensions (a=58.7, 
b=67.4, c=98.9, alpha=78.5, beta=89.9, gamma=89.9) and Phaser finds 8 molecules 
in the AU.  Structure refinement for this scenario is ongoing, and it has yet 
to be seen whether this leads to improvement of the statistical parameters.

I have solved other structures of this protein in complex with different 
substrates.  One such structure was successfully solved in P1 21 1 with two 
molecules in the AU with the unit cell lengths (a2= 47.3, b2=58.9, c2=67.6) and 
angles (alpha2=90, beta2=99.2, gamma2=90).  Interestingly, these dimensions are 
highly similar to those of current structure, although a1 is equivalent to c2, 
and c1 is nearly double a2.

I have run phenix.xtriage on the P2 scaled data, and the Patterson Analysis 
indicates that translational-pseudosymmetry is a posibility in my case.  
Alternatively, phenix.xtriage on the same dataset processed in P1 indicates the 
same translational-pseudosymmetry, but also suggests the presence of a 
pseudomerohedral twin.  I am not sure if this twin simply relates the dataset 
to the higher symmetry P2 group.  The sections of both log files pertaining to 
these analyses are copied below- sorry for all the text!

At this stage, I am wondering what the next steps.  Specificity, is this data 
suspicious for having pseudotranslational symmetry or twinning?  If so, what 
should be my next steps to troubleshoot this problem? Any suggestions would be 
appreciated.

Thanks in advance,

Keith

__

P2 PROCESSED DATA


Patterson analyses: 
--


 Largest Patterson peak with length larger than 15 Angstrom

 Frac. coord.:0.0000.0000.500
 Distance to origin  :   49.477
 Height (origin=100) :   42.327
 p_value(height) :2.073e-04


   The reported p_value has the following meaning:
 The probability that a peak of the specified height
 or larger is found in a Patterson function of a
 macro molecule that does not have any translational
 pseudo symmetry is equal to  2.073e-04.
 p_values smaller than 0.05 might indicate
 weak translational pseudo symmetry, or the self vector of
 a large anomalous scatterer such as Hg, whereas values
 smaller than 1e-3 are a very strong indication for
 the presence of translational pseudo symmetry.



The full list of Patterson peaks is:

  x  y  zheight   p-value(height)
( 0.000, 0.000, 0.500 ) :   42.327   (2.073e-04)
( 0.111, 0.000,-0.476 ) :8.908   (2.385e-01)

 If the observed pseudo translationals are crystallographic
 the following spacegroups and unit cells are possible:

 space groupoperator unit cell of reference setting
   P 1 21 1 (a,b,2*c)   x, y, z+1/2  (67.48, 58.78, 49.48,  90.00, 
101.49, 90.00)

___

P1 PROCESSED DATA

####
##   Twinning Analyses##
####



Usi