Hi Peter, At the moment autosol requires that the datasets all be indexed the same way... All the best, Tom T ________________________________________ From: [email protected] [[email protected]] on behalf of Peter Zwart [[email protected]] Sent: Friday, March 21, 2014 4:30 PM To: Todd Jason Green Cc: [email protected] Subject: Re: [phenixbb] xtriage reindexing suggestions Hi Todd, The symmetry operator that would jumble up I+ and I- is (h,k,l)->(-h,-k,-l) [and symmetry related ones]. Once you provide a dataset that contain anomalous flags, the ASU is expanded such that Friedel mates are considered to be independent reflections (i.e. Friedel symmetry is not part of list of operators). Because of this, the troublesome situation you describe doesn't happen. I wouldn't be surprised if autosol does all these things for you btw, but better confirm with Tom. HTH Peter On 21 March 2014 15:14, Todd Jason Green wrote:

Hello all-

I am trying to merge a few partial Se datasets. I am certain that I will have to reindex one or more to make sure that + and - reflections are not jumbled up. Can someone confirm that xtriage take this into account in suggesting to reindex one dataset in relation to a reference dataset? ie. does it use something like maximizing the anomalous signal in suggesting whether or not reindexing is necessary? I think the below suggests as much but I wanted confirmation:

Automatic re-indexing For each data set supplied (other than the first data set given), all possible re-indexing matrices are derived as described by Grosse-Kunstleve et al. (2005). Each data set isautomatically re-indexed using the matrix that maximizes the correlation of amplitudes

Have a good weekend. -Todd

My Example: Reference analyses The following reindexing operators have been found:

------------------------------------------------- | Operator | Correlation | matches (%) | choice | ------------------------------------------------- | -k,h,l | 0.874 | 70.1 | <--- | | h,k,l | 0.032 | 76.7 | | -------------------------------------------------

If the data is reindexed with operator (-k,h,l), the correlation of the intensities to the reference data is maximized. sx205b_anom.sca This can be done for instance with: phenix.reflection_file_converter sx205b_anom.sca --change_of_basis="-k,h,l"

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- ----------------------------------------------------------------- P.H. Zwart Research 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 SASTBX: http://sastbx.als.lbl.gov ----------------------------------------------------------------- _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Thanks, Tim and others. I think I mixed up two ideas by not clearly stating my situation when I quickly shot out my email on friday. This was compounded by scaling and merging datasets from two crystals with moderate anomalous signal but low completeness together only to find that the anomalous signal completely goes away (ie. looking at the anomalous stats in xtriage). I realize there could be many reasons for this, nonisomorphous data, poor scaling between crystals, etc etc, but i was hoping that i had just missed a convention, and one set needed to be inverted. My case is as follows just to clarify, I have a crystal that is very close to tetragonal but is actually orthorhombic (i think) with a almost equal to b (or worse case scenario heavily twinned primitive monoclinic, but that is another issue altogether). a and b in the orthorhombic setting are each ~180 angstroms, difference in cell axes is ~0.3 angstroms. sometimes, in the initial index the orientation of a and b is swapped. In denzo, I wasn't sure how to switch the axes nor do i know if the switching is necessary without comparing to a reference dataset. So I do actually have to reindex some of the datasets. This is easily identified by xtriage. I wasn't sure if I take two independent dataset with different reference zones, would + and - reflections always be in the same "configuration". I guess it's my own limitation of the algorithm(s) used in the indexing, that were not allowing me to get a grip on the concept. After reading a few more manuals, I trust that all is good. Ie: everything should be indexed on a right handed system, unless you muck around with the wrong setting:

From the CNS tutorial: If the data are correctly indexed on a right handed basis system, the assignment of "F+" and "F-" is such that the experimental electron density map will show amino acids and helices with the correct hand. However, if by accident a left handed system was chosen (for example if the sign of Y SCALE in Denzo was chosen incorrectly), the experimental electron density map will have the wrong hand.

Followed by the original denzo manual: Closer inspection of Figure 5 above should convince you that there are two possible "hands" to conventions. These are grouped as {1,4,5 and 8} and {2,3,6, and 7}. To use an X-ray film analogy, this is equivalent to looking at the film either from the back side(ie. towardsthe X-ray source) or from the crystal side. Denzo handles this abberation by changing the sign of the Y SCALE keyword from positive to negative. If this is not done, then the assignment of the Friedel mates (Bijvoet differences) will be inverted, and thus the hand of the structure will be similarly inverted. As a side note, what is the upper limit on number of Se sites Autosol should be able to find? i have between 200 and 300 depending on the contents of the asymmetric unit. Thanks again for all of the responses. Todd ________________________________________ From: [email protected] [[email protected]] on behalf of Tim Gruene [[email protected]] Sent: Saturday, March 22, 2014 12:08 PM To: [email protected] Subject: Re: [phenixbb] xtriage reindexing suggestions Dear Todd, if you processed your data with XDS, you can provide the REFERENCE_DATA_SET there. Otherwise pointless would also do what you want and in addition aimless lists the anomalous CC of all the pairs if you give each file a separate name in pointless. Best, Tim On 03/21/2014 11:14 PM, Todd Jason Green wrote:

Hello all-

I am trying to merge a few partial Se datasets. I am certain that I will have to reindex one or more to make sure that + and - reflections are not jumbled up. Can someone confirm that xtriage take this into account in suggesting to reindex one dataset in relation to a reference dataset? ie. does it use something like maximizing the anomalous signal in suggesting whether or not reindexing is necessary? I think the below suggests as much but I wanted confirmation:

Automatic re-indexing For each data set supplied (other than the first data set given), all possible re-indexing matrices are derived as described by Grosse-Kunstleve et al. (2005). Each data set isautomatically re-indexed using the matrix that maximizes the correlation of amplitudes

Have a good weekend. -Todd

My Example: Reference analyses The following reindexing operators have been found:

------------------------------------------------- | Operator | Correlation | matches (%) | choice | ------------------------------------------------- | -k,h,l | 0.874 | 70.1 | <--- | | h,k,l | 0.032 | 76.7 | | -------------------------------------------------

If the data is reindexed with operator (-k,h,l), the correlation of the intensities to the reference data is maximized. sx205b_anom.sca This can be done for instance with: phenix.reflection_file_converter sx205b_anom.sca --change_of_basis="-k,h,l"

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen GPG Key ID = A46BEE1A