Mo;lecular replacement can choose any suitable origin so you cant predict.
The easiest way is to run "superpose molecules"
Match all residues to each other.
You should get a rotation matrix -1 0 0 0 0 1 / 0 1 0 with some
translation which sjould be very c;lose tpo components of 0, or 1/2
along all the axes..
ie -x+dx z+dz Y+dy where the dx dy and dz are 0 or 1/2
Eleanor
and that is what you want James Stroud wrote:
Update:
The pounding I am beginning to feel in my head is reminiscent of a
pounding I felt while working in p43212 a while ago. I think this is
an alternate origin issue. Now I just need to figure out how to
redefine the origin.
James
On Apr 16, 2009, at 9:38 PM, James Stroud wrote:
Hello All,
I have two crystals (that I'll call "data set 1" and "data set 2")
that seem to be isomorphic, but y and z are transposed between the
two data sets. Reindexing data set 2 with the operators
h => -h
k => l
l => k
makes the axes match data set 1, but running MR with the previous
data set 2 solution on the newly reindexed data set 2 yields a
solution rotated 180 about z with respect to the data set 1 solution.
What is the operation to reindex such that real space is rotated 180
about z? These are in P212121.
James
