Thanks for taking the time to look at this Peter and Vaheh and other for that matter. I have processed multiple datasets in p1, p222, r3, and p23. I'll try to compile a something from the log files to submit for further questions that I have. I think it is possible that some of my datasets could be twinned, I'm certainly looking into it. xtriage doesn't seem to think so for the couple of datasets that i have run through it so far, although in some of the lower symmetry sg there are a few twin laws which indicate so, xtriage seems to attribute it to orientation of ncs. I'm also doing molecular replacement in each of the space groups. Also in some of the datasets, I have alternating lattice lines of heavy and light reflections. Thusly without carefully selecting spots and the light reflections will not be chosen and the wrong lattice will be assigned(ie. the lower density spots get skipped). It goes without saying that some of the datasets are not optimal but its what i have at the moment. One of the first red flags for me in regard to it not being cubic (at least one of the datasets), was that if i processed two seperate wedges of data from the same crystal not using the same indexing then tried to merge the datasets together the r-merge would be .4+. If i used xtriage to determine an appropriate reindexing operator then merged the data, the r-merge would be .08. I didn't think alternate indexing was possible for the cubic space group (other than for the assignment of +/- reflection). So even though in my case it is Se-met data, i didn't think that the difference of 30+ percent between alternative indexings could be possible. I thought this might be indicating that I didn't have a true p23 space group. Is my thinking correct here? Also in some of the datasets, I have alternating lattice lines of heavy and light reflections. Thusly without carefully selecting spots, the light reflections will not be chosen and the wrong lattice will be assigned. It goes without saying that some of the datasets are not optimal but its what i have at the moment. Are the rows of light reflections typically indicative of pseudo-symmetry? I believe that I will have 20 copies of a hetero-complex in the cubic setting and 120 in orthorhombic which should have i believe a packing similar to the cubic crystal. thanks in advance- Todd -----Original Message----- From: [email protected] on behalf of Peter Zwart Sent: Fri 8/10/2007 5:24 PM To: PHENIX user mailing list Subject: Re: [phenixbb] conversion from h3 to r3 with phenix apps Hi Todd, This table tells me that the data in P1 doesn't merge very well and that limits for what is reasonable are maybe set to tight for your case. Try using the keyword tanh_location=0.11 and see what happens. Just looking at the table though, there seem to be a number of options to be considered: P1: one unused symmetry operator present with R value equal 10%. Not a good sign, sg is too low P432: ONe operator has an R value of 47%. Not a good sign, sg is too high A couple of flavours of R3. | R 3 :h (x+z,y-z,-x+y+z) | 0.148 | 0.148 | 0.256 | 0.103 | | R 3 :h (-x+y+z,x+z,y-z) | 0.153 | 0.153 | 0.256 | 0.103 | | R 3 :h (y-z,-x+y+z,x+z) | 0.164 | 0.164 | 0.256 | 0.103 | | R 3 :R | 0.169 | 0.169 | 0.256 | 0.103 | From the last column it is clearly seen that an operator is present with a R-value equal to 10%, but is not used in the point group. The point group might be too low. The flavours of R32 don't work, as a operator is used that has a high R value: | R 3 2 :h (-x+y+z,x+z,y-z) | 0.484 | 0.489 | 0.109 | 0.103 | R 3 2 :h (y-z,-x+y+z,x+z) | 0.486 | 0.492 | 0.109 | 0.103 These options are nice(ish): | P 2 3 | 0.134 | 0.169 | 0.478 | 0.478 | | | P 2 2 2 | 0.109 | 0.110 | 0.481 | 0.478 | | | P 4 2 2 (b,c,a) | 0.181 | 0.486 | 0.480 | 0.478 | | Do you have any indication that your data is twinned? Peter Green, Todd wrote:
I'm actually not sure that it's not orthorhombic. I know looking at the cell lengths and angles makes it very inviting to the cubic cell and i have been working with that as a possibility. i do have a real good suspicion that I have a pseudo-cubic cell. I can do molecular replacement in the cubic space group but am missing some density that I think should be there. I wanted to do molecular replacement in the rhombohedral setting and ultimately the lower symmetry space group too and see if I pick that density up. This is why I wanted to know how to convert the h3 to r3 setting. the h3 cell is a nightmare for my computer!
as a side note not all of the crystals, process in the primitive lattice, but all process as rhombohedral. That probably means that I have a two different types of crystals more than anything else, etc etc etc. cross-xtal averaging is another reason to i wanted to figure out the setting difference.
I have previously run one of my datasets processed in p1 thru Xtriage, it said:
Exploring higher metric symmetry
The point group of data as dictated by the space group is P 1 the point group in the niggli setting is P 1 The point group of the lattice is P 4 3 2 A summary of R values for various possible point groups follow.
------------------------------------------------------------------------------------------------ | Point group | mean R_used | max R_used | mean R_unused | min R_unused | choice | ------------------------------------------------------------------------------------------------ | P 1 | None | None | 0.362 | 0.103 | <--- | | P 2 3 | 0.134 | 0.169 | 0.478 | 0.478 | | | P 4 | 0.152 | 0.488 | 0.377 | 0.164 | | | C 1 2 1 (z,x+y,-x+y) | 0.489 | 0.489 | 0.337 | 0.103 | | | P 4 2 2 (c,a,b) | 0.266 | 0.481 | 0.481 | 0.478 | | | P 1 1 2 | 0.110 | 0.110 | 0.221 | 0.103 | | | P 2 1 1 | 0.107 | 0.107 | 0.290 | 0.110 | | | R 3 2 :R | 0.481 | 0.483 | 0.109 | 0.103 | | | C 1 2 1 (x+y,-x+y,z) | 0.492 | 0.492 | 0.269 | 0.103 | | | P 2 2 2 | 0.109 | 0.110 | 0.481 | 0.478 | | | C 2 2 2 (x-y,x+y,z) | 0.190 | 0.478 | 0.339 | 0.103 | | | P 4 (c,a,b) | 0.454 | 0.476 | 0.272 | 0.110 | | | C 1 2 1 (x+y,z,x-y) | 0.483 | 0.483 | 0.351 | 0.103 | | | C 1 2 1 (-x+y,z,x+y) | 0.486 | 0.486 | 0.351 | 0.103 | | | P 1 2 1 | 0.103 | 0.103 | 0.343 | 0.110 | | | R 3 2 :h (x+z,y-z,-x+y+z) | 0.488 | 0.492 | 0.109 | 0.103 | | | P 4 2 2 (b,c,a) | 0.181 | 0.486 | 0.480 | 0.478 | | | R 3 :h (x+z,y-z,-x+y+z) | 0.148 | 0.148 | 0.256 | 0.103 | | | C 2 2 2 (z,x-y,x+y) | 0.426 | 0.489 | 0.309 | 0.110 | | | R 3 2 :h (-x+y+z,x+z,y-z) | 0.484 | 0.489 | 0.109 | 0.103 | | | C 2 2 2 (-x+y,z,x+y) | 0.386 | 0.486 | 0.327 | 0.110 | | | P 4 2 2 | 0.243 | 0.492 | 0.386 | 0.169 | | | R 3 :h (-x+y+z,x+z,y-z) | 0.153 | 0.153 | 0.256 | 0.103 | | | R 3 :h (y-z,-x+y+z,x+z) | 0.164 | 0.164 | 0.256 | 0.103 | | | P 4 3 2 | 0.362 | 0.492 | None | None | | | C 1 2 1 (x-y,x+y,z) | 0.478 | 0.478 | 0.269 | 0.103 | | | R 3 2 :h (y-z,-x+y+z,x+z) | 0.486 | 0.492 | 0.109 | 0.103 | | | R 3 :R | 0.169 | 0.169 | 0.256 | 0.103 | | | P 4 (b,c,a) | 0.419 | 0.475 | 0.290 | 0.110 | | | C 1 2 1 (z,x-y,x+y) | 0.481 | 0.481 | 0.337 | 0.103 | | ------------------------------------------------------------------------------------------------
R_used: mean and maximum R value for symmetry operators *used* in this point group R_unused: mean and minimum R value for symmetry operators *not used* in this point group The likely point group of the data is: P 1
Possible space groups in this point groups are: Unit cell: (375.144, 375.711, 377.723, 90.002, 90.034, 90.094) Space group: P 1 (No. 1)
what does this say to you?
Thanks in advance- Todd
-----Original Message----- From: [email protected] on behalf of Peter Zwart Sent: Fri 8/10/2007 4:35 PM To: PHENIX user mailing list Subject: Re: [phenixbb] conversion from h3 to r3 with phenix apps
Hi Todd,
Are you sure this is not cubic? You could run xtriage and find out relatively easely.
I suggest you give phenix.xmanip a try for reindexing or try to use the following one-liner:
iotbx.reflection_file_converter data.sca --change-of-basis=to_niggli_cell --sca=niggli.sca
Cheers
Peter
Green, Todd wrote:
Hello all,
I have what i believe to be a rhombohedral crystal that has an insanely large cell with the hexagonal setting:
533.026 533.026 652.887 90.000 90.000 120.000
and a modestly large cell comparatively with the rhombohedral setting:
377 377 377 90 90 90
I should be able to easily reindex to the smaller cell in scalepack but for some reason i'm not getting it to work correctly. Rather than struggle further on a friday afternoon, i figured that i'd give a phenix app a try. i assume xtriage can do this, can someone point me in the correct direction?
thanks in advance- Todd
------------------------------------------------------------------------
_______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
This email was scanned with Mcafee's Anti-Virus appliance, but this is no guarantee that no virus exists. You are asked to make sure you have virus protection and that it is up to date.
------------------------------------------------------------------------
_______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
This email was scanned with Mcafee's Anti-Virus appliance, but this is no guarantee that no virus exists. You are asked to make sure you have virus protection and that it is up to date.