[phenixbb] pseudo translational symmetry

Peter Zwart phzwart at gmail.com
Mon Jun 22 22:24:34 PDT 2009

Hi Youssef,

Some (leading) questions

1. How do you images look?

2. For your structure, what is the quality of the density for each
monomer, are both the same or is one monomer radically different from

3. How certain are you of your space group?

Some suggestions in order of decreasing sanity

Strategy A:

1. Expand your data set to P1

phenix.reflection_file_converter mydata.sca --expand_to_p1

 (or reprocess in p1)

2. Solve this structure with the model you have (using MR or expanding
the structure to P1)

3. Refine this structure with phenix.refine (it chooses the free set
with the highest symmetry in mind)

4. Run xtriage on both the p1 data and model at the same time
(phenix.xtriage data.sca reference.structure.file=model.pdb). The RvsR
statistics can indicate the presence of higher, pseudo symmetry,
twinning or a combination of these problems.

Strategy B:

Try switching target functions (use LS instead of ML). The presence of
pseudo translation function is known to confuse the likelihood model
in molecular replacement, and I wouldn't be surprised if it could have
a similar effect in refinement.

Alternatively, cheat the system by refining alternatively against weak
and strong data until you have pushed the system down the deep

Strategy C:

Shift the model you have by (0,0.25,0) and see how that refines

Strategy D:

"Grow better crystals"



2009/6/22 youssef ben ammar <ybenammar at gmail.com>:
> Hi all,
> I recently solved the structure of a protein containing 395aa using phenix
> autosol. After autoBuild I got about 65% of the sequence docked with 2
> monomers in the ASU.
> After several refinement cycles the map has been improved and I build
> manually the remaining amino acids. But the problem is that R and Rfree
> didn't decrease below 0.28 and 0.33 respectively. I tried with NCS
> restraints, TLS but without success.
> When I revised phenix autosol logs carefully, I found that in xtriage the
> analyses of the Patterson function reveals a significant off-origin peak
> that is 41.59 % of the origin peak, indicating pseudo translational symmetry
> and no twin laws were found.
> The space group is P212121 (36.898, 59.45, 393.908, 90, 90, 90). The basic
> statistics suggested 1 copy in the ASU (but the solution gave 2 copies per
> ASU !!!). The same analysis done by ccp4i gave the same solvent content and
> math. coef. in p21212 but with 2 copies per ASU.
> In xtriage I found this suggestion: If the observed pseudo translationals
> are crystallographic the following spacegroups and unit cells are possible:
> space group  P 21 21 2 (b-1/4,2*c,a)          operator    x, y+1/2, z
> unit cell of reference setting  (393.91, 36.90, 29.73,  90.00, 90.00, 90.00)
> I am really stucked with the refinement and don't know how to deal with this
> pseudo translation.
> Any help or suggestions are welcome.
> Thank you.
> Youssef
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb

P.H. Zwart
Beamline Scientist
Berkeley Center for Structural Biology
Lawrence Berkeley National Laboratories
1 Cyclotron Road, Berkeley, CA-94703, USA
Cell: 510 289 9246
BCSB:     http://bcsb.als.lbl.gov
PHENIX: http://www.phenix-online.org
CCTBX:  http://cctbx.sf.net

More information about the phenixbb mailing list