Installing SBGrid Software
Using the SBGrid Environment
Support for Site Administrators
Hardware Support Notes
Support for Developers
The first follow-up step for a search model resulting from WSMR is to repeat MR locally.
To download one of the search models; first click the "Load Top 100 Hits" button, click one of the yellow circles in the resulting plot, and then use the download link for the model.
A typical input deck (
MR01.in for this example) for
phaser would look something like this (with the paths to your input
PDB files adjusted appropriately):
ROOT MR01 MODE MR_AUTO HKLIn reflections.mtz LABIn F=FP SIGF=SIGFP ENSEmble SearchModel PDB model.pdb IDENtity 0.3 COMPosition PERCENTAGE 0.50 RESOlution HIGH 3.1 SEARch ENSEmble SearchModel NUM 1 SGALternative SELECT ALL
This can be run with versions of
phaser distributed with
phaser < MR01.in > MR01.out or
phenix.phaser < MR01.in > MR01.out respectively). The statistics (LLG and Z-scores) from local runs may differ due to differences in
SGALternative card is only necessary if you're checking alternative space-groups.
WSMR searches for a single copy of the search model - however there may be multiple copies of the model in the ASU. If so, the next step is the increase the
NUM value for the
SEARch card. When in doubt, it's better to set this too hight rather than too low -
phaser will stop when unable to locate additional copies.
Once MR has been exhausted (no more copies of the search model can be placed), one possible next step is to use an autobuilder for an initial attempt to correct the sequence of the model.
phenix.autobuild reflections.mtz MR_Results.pdb project.fasta nproc=4 > AutoBuild01.out
The input model (
MR_Results.pdb) for the autobuilder is the final result modeles from the initial or expanded MR search.
MTZ file should be the original amplitudes (or intensities) used for searching, not the results from
phaser. If phaser was run with the
SGALternative card, make sure the spacegroup of this
MTZ matches that of the search solutions.
nproc=4 option should only be used if there are four (or more) available cores - when using the
phenix.autobuild can't take advantage of more than 4 processors.