 |
AMoRe |
The idea of Molecular Replacement is to build a tentative crystal structure using known molecular models, similar to the actual molecules that constitute the crystal, in order to start model building or refinement. The problem is to determine the positions of the models within the crystal cell, which is ultimately done by comparing observed and calculated structure factors, for selected positions of the independent molecules within the cell. In AMoRe the comparison involves essentially the correlation coefficient in terms of amplitudes.
Parallel use of AMoRe would best be an examination of different resolution ranges, B-factor modulation, and model testing.
Author: Jorge Navaza
How to use
Source the initialization file by typing source /usr/local/AMoRe/login
Create the input pdb models xyz1.d, xyz2.d, ... xyzN.d for N models. Include the format line as the first line of the file (only the atom type, x, y, and z coordinates, and occupancy and temperature factors are read):
xyz1.d:
FORMAT (12X,A4,14X,3F8.3,6X,F6.2)
ATOM 1 N LYS 1 -11.460 31.361 22.771 1.00 20.48 N
...
Then create the input hkl file (h, k, l indices and the Fobs, sigma is not read), also including the format line:
hkl.d:
FORMAT (3I4,F12.2)
0 0 6 895.10 38.40
0 0 8 507.50 15.70
...
Create a new directory for the work, and cd to it. Type csh $AMORE/setup to create a set of working directories and scripts. cd to directory d, copy the xyzN.d and hkl.d files to this directory, and edit the data.d file:
** D44HEL **
103.9 38.7 34.0 90. 100.6 90.
x,y,z * -x,y,-z * 1/2,1/2,0 * end
0 ; orthogonalising code
95. 0. ; % reflections, b-add
15. 3.5 ; resolution range
1 1 ; NTYP, (nmol(n),n=1,NTYP)
>>> logicalunits 1-9 and 90-95 are used by AMoRe.
Once all these files are created and are in a single directory, the AMoRe job can be launched onto the cluster from the new directory.
Here is a sample qsub script:
#!/bin/csh
#PBS -N amore
#PBS -e amore.err
#PBS -o amore.log
cd $PBS_O_WORKDIR
source /usr/local/AMoRe/amore_init
csh ./e/job dato
csh ./job >& log
Parallel use
Because an individual AMoRe run lasts less than 15 minutes, the best use of AMoRe is to screen various parameters, such as resolution ranges, search models, and temperature factors. The best way to do this is to create a series of directories with either the entire contents duplicated, or at least symbolic links to a central directory (for standard parameters and inputs). Then modify the files within each directory and submit the whole lot with a series of qsub commands, or a single swarm command.
Output
In the directory ./o, a series of files is created, detailing first the rotation peaks and their scores, followed by the translation peaks and their scores for the combinations of models. The standard output has a summary of the solution. More details can be found in the tutorial.
More information and finer details
