Dear Alex,

Could you clarify what you would like to see when you check all space groups?  Phaser (and some other programs) can check all the space groups with the same point group symmetry, i.e. all the space groups that are compatible with the symmetry used in processing and merging the data.  (Or at least the space groups without inversion operators that chiral molecules like proteins can adopt.)  Any wider choice of space groups would require you to merge the data either with lower or higher symmetry, but then you should be looking at other indicators (twinning tests, extra symmetry in the diffraction data) to address those uncertainties.

Actually, Phaser will now allow you to specify a lower symmetry subgroup of the point group in which the data have been provided.  This is useful if it looks like the data have been merged in too high symmetry because a twin operator has mistakenly been interpreted as part of the space group symmetry.  In this case, Phaser expands the data to the lower symmetry.  But you can only test one subgroup at a time, and the choice has to be made manually.

Best wishes,

Randy

On 24 Apr 2016, at 20:38, Alex Lee <[log in to unmask]> wrote:

Dear Stefan,

Thank you very much for the information! For Balbes "check full space group", is it going to check every possible space group or is it like "Run Phaser with all alternative space groups" in Phaser MR which seems to only have a few space group choices (like SG P312, the lists of alternative space groups in Phaser MR are P3112 and P3212)?  

Another question is if I have multiple chains in a protein complex, for example if I need to search two or three different chains, can Balbes handle it after I put all chain sequence into a file?   I tried MrBUMP in CCP4online for a three chain MR (two chains are the same for a homodimer and the third chain is from the homodimer's partner) and MrBUMP only generate model from one chain (one of the homodimer) and do MR,  even-though I put three chains into a file  in the sequence input. 





On Sun, Apr 24, 2016 at 6:08 AM, Stefan Arold <[log in to unmask]> wrote:
Dear Alex,
in case you have other things to do and can wait for a few hours: just upload your mtz and paste your sequence into Balbes (one of the programs found at http://www.ccp4.ac.uk/ccp4online). Click 'check full space group' and 'run arp/warp' as needed. 
all the best,
Stefan

Stefan T Arold, PhD

Associate Professor

Division of Biological and Environmental Sciences and Engineering

Computational Bioscience Research Center (CBRC)
King Abdullah University of Science and Technology (KAUST)

Thuwal 23955-6900
Bldg 2, Level 4, Room 4275
Kingdom of Saudi Arabia

Office: +966 (0)12 808 2557

Mobile: +966 (0)5 4470 0432

Em: [log in to unmask]



On 23 April 2016 at 00:52, Alex Lee <[log in to unmask]> wrote:
Hi All,

I really appreciate all of your suggestions! For my simple case, "allow search with alternative ensembles for a single component of ASU" in Phaser would work. For more difficult case, I have learned to generate superimposed ensembles following your suggestions about using Ensembler or MrBUMP.





On Fri, Apr 22, 2016 at 2:30 AM, Gabor Bunkoczi <[log in to unmask]> wrote:
Hi Alex,

if you want Phaser to select the best of a set of alternative models, you need to use the "Add ensemble" button below the one you found, define all six models (they do not have to be superimposed), and then turn on "Allow search with alternative ensembles for a single component of ASU" in "Additional search parameters", and specify all ensembles you defined in the search window. Phaser will automatically search with each model, and select the best based on the LLG-score.

Although it is not relevant to your problem, you can use Ensembler (available through the "Model generation" submenu of "Molecular replacement") to create a set of superposed models. Ensembler has an iterative procedure to automatically weight up structurally conserved regions in your models, which often improves model quality compared to an unweighted superposition.

BW, Gabor


On 2016-04-22 08:06, Alex Lee wrote:
Hi All,

I am using Phaser MR to solve structure with a single protein chain
(around 70 amino acid length). There are 6 homolog candidate models
with sequence identity from 90% to 70%. I could run 6 jobs separately
with each model as an ensemble to see which model gives best solution.
But I am thinking is there a more convenient approach to put the 6
candidate models in one job and let Phaser MR compare which one gives
the best solution?

After I did some Internet search, I do not know if this is called "add
superimposed pdb file to the ensemble" in Phaser MR GUI (in CCP4),
where I just add the six conadidate models (PDBs) before starting run.
I also notice a program called Superpose in CCP4, I do not know if I
need this program to superimpose my 6 models first and then add the
superimposed pdb (in this case I guess it's the final one pdb output
after superimposed 6 models) in to Phaser MR before starting run?

Thanks in advance.

--
##################################################

     Dr Gabor Bunkoczi

     Cambridge Institute for Medical Research
      Wellcome Trust/MRC Building
     Addenbrooke's Hospital
     Hills Road
     Cambridge CB2 0XY
##################################################




------
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: [log in to unmask]
Cambridge CB2 0XY, U.K.                       www-structmed.cimr.cam.ac.uk