[phenixbb] xtriage reindexing suggestions

Todd Jason Green tgreen at uab.edu
Mon Mar 24 08:32:20 PDT 2014


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:



More information about the phenixbb mailing list