Previous Thread
Next Thread
Print Thread
image transformation
#10275 04/17/06 06:31 PM
Joined: Aug 2005
Posts: 28
H
Forum Member
OP Offline
Forum Member
H
Joined: Aug 2005
Posts: 28
hello all,

I have a question about the image transformation.
I define a orth simulaton box with x,y,z length of 45,45, and 60 A respectively.
then, I use crystal build using p1 symmetrical operation.

the output is as follows.
CHARMM> crystal define orth 45 45 60 90 90 90
Crystal Parameters : Crystal Type = ORTH
A = 45.00000 B = 45.00000 C = 60.00000
Alpha = 90.00000 Beta = 90.00000 Gamma = 90.00000

CHARMM> crystal build cutoff 20
XBUILD> Building all transformations with a minimum atom-atom
contact distance of less than 20.00 Angstroms.

Range of Grid Search for Transformation 1 :
Lattice Vector A -2 TO 2
Lattice Vector B -2 TO 2
Lattice Vector C -1 TO 1

The number of transformations generated = 6
Number Symop A B C Distance

1 1 -1 -1 0 18.2744
2 1 -1 0 0 16.6973
3 1 0 -1 0 15.5381
4 1 0 1 0 15.5381
5 1 1 1 0 18.2744
6 1 1 0 0 16.6973
THERE ARE NO ROTATIONS FOR THIS TRANSFORMATION SET
6 Transformations have been processed.

please correct me if my understanding is wrong.
For the transformation, for example, in the second case (-1,0,0), why the distance is 16.6973, not full box x-length of 45 A. I assumed (-1,0,0) means the image box was placed along the negative direction of x axis.

how charmm get the transformation distance?

Re: image transformation
hunulolo #10276 04/17/06 06:48 PM
Joined: Sep 2003
Posts: 8,470
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,470
The 'Distance' column identifies the shortest inter-atom pair distance which triggered the addition of the transformation. The output suggests that your unit cell shape still lots of empty space, and that consequently the CUTOFF for CRYSTAL BUILD is not large enough. There are way too many missing transformations (min 26 needed for this lattice type and symmetry), and the distances listed are only a little less than 20. For this case, you should increase CUTOFF from 20 to at least 30 (1/2 the longest edge), to get all of the needed transformations.


Rick Venable
computational chemist

Re: image transformation
rmv #10277 04/17/06 07:07 PM
Joined: Aug 2005
Posts: 28
H
Forum Member
OP Offline
Forum Member
H
Joined: Aug 2005
Posts: 28
Can charmm write the image coordiates with original coordinate in crd or pdb format? so, I can check whether image transformation is right?

Re: image transformation
hunulolo #10278 04/17/06 08:38 PM
Joined: Sep 2003
Posts: 8,470
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,470
update imall cutim 30
open unit 12 write card name atomimg.pdb
write coor pdb unit 12 image
* real atoms and any image atoms within CUTIM
*


I've assumed a "mostly empty" unit cell, as indicated by your first post; only image atoms within CUTIM of a real atom will be written to the output coord file.


Rick Venable
computational chemist

Re: image transformation
rmv #10279 04/18/06 06:26 PM
Joined: Oct 2004
Posts: 22
L
leo Offline
Forum Member
Offline
Forum Member
L
Joined: Oct 2004
Posts: 22
If my understanding is correct, the number of transformation needed depends on the lattice type and symmetry. So how could we know whether it is enough or not? Is there any criteria we can follow except that we write its image coordinates to a pdb file?

Thanks,

-Leo

Re: image transformation
leo #10280 04/18/06 07:11 PM
Joined: Sep 2003
Posts: 8,470
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,470
For ORTH, TETR, and CUBIC lattices with P1 symmetry and edge lengths greater than 2*CTOFNB, 26 transformations are generally enough; this puts the 'box' of real atoms in the center of a 3x3x3 grid.

For RHDO and OCTA with P1 symmetry, I believe the minimum number may be the number of faces, but I haven't verified that.

It should be noted that CRYSTAL BUILD works best when the volume of the unit cell is mostly filled with atoms in a configuration that is a reasonable match for the lattice shape. One reason is that the CUTOFF used can be closer to the intended CUTIM value, and one can still expect to get enough transformations defined. Problems can arise, however, for a small protein in a large box, with lots of empty space; the CUTOFF must be much larger, up to 1/2 the longest box diagonal, in order to get all the transformations defined correctly.

Once the unit cell (or asymmetric unit) is fully packed with atoms in the correct shape, one can vary the CUTOFF used with CRYSTAL BUILD (e.g. start with CUTNB, and increase in 1 A incr) and use ENERGY to evaluate the result; if the potential changes when CUTOFF is increased, the lower value did not define enough transformations.


Rick Venable
computational chemist


Moderated by  lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 5.6.33-0+deb8u1 Page Time: 0.007s Queries: 26 (0.003s) Memory: 0.9337 MB (Peak: 1.0313 MB) Data Comp: Off Server Time: 2020-07-14 07:27:04 UTC
Valid HTML 5 and Valid CSS