reindexing monoclinic data

From: Gregory Bowman
Date: 26 August 2011 16:55

Hi all,


We have several primitive monoclinic datasets for the same protein with various ligands, with essentially the same unit cell parameters. We would like to have these with the molecules/density oriented the same way for easy comparison, but as chance would have it, some have effectively the opposite "k" index, which of course puts these molecules/density (relatively) upsidedown. I was wondering how people typically deal with this. I found what I believe to be the answer of reindexing monoclinic h, k, l to  -h, -k, h+l to keep the system right handed and flip k. For this it seems that SFTOOLS would be appropriate? Is this reindexing commonly done at the stages of integration (altering rotx roty rotz in HKL2000) or scaling?

Thanks,
Greg

----------
From: Tim Gruene



Dear Greg,

with XDS I normally set "REFERENCE_DATA_SET" to the first one indexed /
integrated in order to maintain consistent indexing between data sets.
Not sure whether similar options are available in other integration
programs, but if I remember correctly, pointless also offers reindexing.

Cheers, Tim

- --
- --
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen



----------
From: Harry


Hi

Yes, in the CCP4 world, pointless is the program for you.

Harry
--
Dr Harry Powell, MRC Laboratory of Molecular Biology, MRC Centre, Hills Road, Cambridge, CB2 0QH

----------
From: Edward A. Berry

If you want to redo in HKL, look at the denzo/scalepack manual,
scalepack scenario 5, reindexing.
I think your case is the first example, "switch a and c in p21"

HKL MATRIX
       0 0 1   [reindexing matrix: h -> l]
       0 -1 0  [this is -1 to keep determinant = 1]
       1 0 0   [l -> h]

If you just want to reindex the data, you can use the previous scalepack
output (.sca) as input. If you rescale the denzo output files (.x)
with this HKL matrix, then in the log file, just below the unique
occurance of the word "mosaicity" in the file, is a line (or several
lines if "fit batch") which gives the post-refined missetting angles
and cell param, which can be edited into the corresponding lines of
integ.inp if you want to re-integrate in the re-indexed setting.
(I don't know if there is any good reason to do so, however).

----------
From: Bosch, Juergen

Hi Greg,

you could also CAD them into one huge mtzfile with different labels e.g. FP_Lig1, FP_Lig2 etc. Then they would be the way you need them for direct comparison in Coot, Pymol or whatever program you wish to use.

Jürgen

......................

Jürgen Bosch

Johns Hopkins University

Bloomberg School of Public Health

Department of Biochemistry & Molecular Biology

Johns Hopkins Malaria Research Institute

615 North Wolfe Street, W8708

Baltimore, MD 21205

----------
From: Edward A. Berry

Gregory Bowman wrote:

Yes, but I don't actually want to swap a and c (convention that a is shorter than c), but instead flip k and keep h and l the same. Incidentally, it is not immediately obvious to me why in matrix I cited below that the new l is h+l:
-1 0 0
0 -1 0
1 0 1

Greg

Because -a and c would give acute beta angle, by convention should be obtuse?



----------
From: Edward A. Berry


Sorry for adding confusion- of course in reciprocal space beta* should be
acute. The figures should have been labeled a, c not h,L.