A number of superposition programs allow to superimpose specified atoms (such as CA).
Once you get the operator, comparing two different operators is not a job
for a conventional superposition program, since you are superimposing a
line on a line which has the extra degree of freedom- rotation about the line.
If you express the operator in spherical polar coordinates, which the superposition
program may provide or you can get from the matrix using ccp4 "rotmat",
you should be able to work out the relation between the axes.
Using the Uppsala sofware factory programs:
#This is superimposing parts of chains C, A, B on P, N, O (and vice versa - it's proper 2-fold)
lsqman -b <<eof
chain_mode original
re m1 cbc596.pdb
exp m1
C20-370 P20-370 A20-200 A250-400 B30-200 B250-400 N20-200 N250-400 O30-200 O250-400
m1
P20 C20 N20 N250 O30 O250 A20 A250 B30 B250
save m1 m1 ncsasc01.odb
quit
eof
This prints the operator and saves it in ncsasc01.odb*
----------------
- The 2010 atoms have an RMS distance of 0.130 A
- RMS delta B = 15.367 A2
- Corr. coeff. = 0.6601
- Rotation : -0.834671 -0.550747 -0.001321
- -0.550747 0.834660 0.004400
- -0.001321 0.004400 -0.999989
- Translation : 129.384 38.428 171.594
-------------------
Now use mama in an off-label way to convert to polar coordinates:
mama
overlap ncs ncsasc01.odb
gives:
---------------------------
- RT-OP 1 = -0.8346710 -0.5507473 -0.0013208 129.384
- -0.5507473 0.8346604 0.0044000 38.428
- -0.0013208 0.0044000 -0.9999894 171.594
- Determinant of rotation matrix 1.000000
- Column-vector products (12,13,23) 0.000000 0.000000 0.000000
- Crowther Alpha Beta Gamma 106.709 179.736 73.291
- ***Spherical polars Omega Phi Chi 90.132 -73.291 180.000***
- Direction cosines of rotation axis 0.287514 -0.957774 -0.002303
- X-PLOR polars Phi Psi Kappa *undefined* *undefined* 180.000
- Lattmann Theta+ Theta2 Theta- -180.000 179.736 -146.582
- Rotation angle 180.000
-------------------------
* if you use the .odb file outside of the USF software, be aware the matrix is the transpose
(or more accurately it is written by columns). In ccp4 this is taken care of withthe keyword "odb".
On 06/19/2015 09:07 AM, Tobias Beck wrote:
> Dear all,
>
> I have a PDB file that contains NCS in the asymmetric unit, probably point group D3.
>
> 1.) What program is recommended for determining the symmetry operators from PDB coordinates? I found findncs, but this uses only heavy atom coordinates (I could probably use just the sulfurs from the PDB as a work around).
>
> 2.) Then I would like to compare the PDB file to a related structure. Here I would like to align the symmetry operators determined above with symmetry elements found in a different space group, for example align the twofold axis from NCS with a twofold axis in found in a particular space group.
> What is a good way to go about this?
>
> I am aware that NCS is used in programs as restraints during refinement, but here I am interested in obtaining the NCS symmetry operators and aligning them to symmetry elements present in a new space group. Maybe I am overlooking an obvious solution.
>
> Any help is greatly appreciated.
>
> Thanks and best wishes, Tobias.
> --
> ___________________________________________
>
> Dr. Tobias Beck
> - independent group leader -
> RWTH Aachen University
> Institute of Inorganic Chemistry
> Landoltweg 1, office: 304N
> 52056 Aachen, Germany
> phone: +49-241-80-90057
> fax: +49-241-80-99003
> web: http://www.ac.rwth-aachen.de/extern/beck/
> ___________________________________________
>
|
