Home | RAPPERtk | People | RamPage | Download | Docs | FAQ | Search | RSS feed | Contact Us Bioinformatics Group
University of Cambridge 
Protein Structure Modelling  

RAPPER FAQ

rapper ca-trace --pdb my.pdb --map my.map --mainchain-restraint-threshold 2 --chi-squared-electron-density-scoring true --cryst-d-high 3 --edm-rebuild-poor-regions true --edm-poor-region-threshold 0.9 --edm-fit true --use-edm-filters true --edm-sidechain-min-sigma 0.5 --edm-mainchain-min-sigma 0.5 --write-user-remarks true

The first basic question is, did RAPPER run to completion? If RAPPER is still running, then you won't see any output files because RAPPER writes out all of the model PDB files at the end of the program. A related concern is that the electron density parameters are very strict. You're asking for a very good fit to the map (an edm-poor-region-threshold of 0.9) while generating modeling by fitting sidechains (edm-fit is true) and using iterative model filters (use-edm-filters is true). Consequently, RAPPER will probably rebuild a lot of your structure, and will take a long time to do it because the calculations are intense with fit and filters enables.

The second issue, is are you looking in the correct place of the output files? RAPPER writes out all of its files to a subdirectory in your current directory called TESTRUNS. Why TESTRUNS? History -- the first RAPPER implementation was running tests, and the name stuck. Look in the directory where you are running RAPPER with 'ls'. You should see something like:

drwxrwxrwx 9 mdeprist mdeprist 306 Aug 9 2004 TESTRUNS

Look inside of test runs with 'ls TESTRUNS' and you should see something like:

[mark-a-depristos-power-mac-g5 LOOP]# ls -ltr TESTRUNS/ -rw-rw-rw- 1 mdeprist mdeprist 4342 Sep 14 14:39 native.pdb -rw-rw-rw- 1 mdeprist mdeprist 384 Sep 14 14:39 models.dat -rw-rw-rw- 1 mdeprist mdeprist 2915 Sep 14 14:39 looptest.pdb -rw-rw-rw- 1 mdeprist mdeprist 2915 Sep 14 14:39 looptest-best.pdb -rw-rw-rw- 1 mdeprist mdeprist 76378 Sep 14 14:39 framework.pdb -rw-rw-rw- 1 mdeprist mdeprist 504 Sep 14 14:39 benchmark.dat -rw-rw-rw- 1 mdeprist mdeprist 27227 Sep 14 14:40 run-parameters.xml

The issue is that the Mandrake linux tcsh appears to be different from the standard, and that it doesn't like the tcsh script syntax. A simple work around is to specialize your rapper script so that it doesn't try to figure out your system architecture but simply runs the distribution for linux:

#!/bin/tcsh

setenv arch Linux # one of Darwin FreeBSD or Linux
setenv proc i386 # one of i386 or powerpc
setenv EXEC $0:h/distrib/rapper_${arch}_${proc}
setenv PARAMS $0:h/distrib/params_${arch}_${proc}.xml
setenv MODE $1
shift
$EXEC $PARAMS $MODE --rapper-dir $0:h $argv

This should work for you, but only on your Mardrake linux system.

*** The following errors have occurred during rappermc.py execution: (1) The following command exited with status [34304] <> 0: /programs/i386-linux/rapper/rapper ca-trace [BUNCH OF RAPPER ARGS] rapper_Linux_i386: atomconv.cpp:897: int expected_max_num_atoms(residue_type): Assertion `is_protein(restype)' failed.

Your PDB file probably includes non-protein atoms labeled as ATOM records and included in the residues for RAPPER to rebuild. Currently RAPPER can only sample protein conformations, and exits not so gracefully if you try to build DNA, RNA, ligands. The reason that RAPPER is trying to rebuild your DNA, RNA etc is because the default.cfg chain-id parameter is set to '-', meaning all chains. The work-around is to make the DNA, RNA, etc. into HETATMs and put them at the end of the file, so RAPPER won't think that they are part of your protein.

Yes, RAPPER is more than happy to operate on multiple chains. It rebuilds each part of the chain independently, so it knows not to try to connect the end of chain A with the start of B, for example. In the standard defaults.cfg file, rappermc.py builds all chains in the PDB file (chain-id: -). You can change it to operate only a single chain (chain-id: A, for example). If you have some chains that are not proteins (i.e., DNA or ligand), it's best to make them HETATMs so that RAPPER will completely ignore them.

ERROR [Internal RAPPER Bug]: in file params.cpp at line 875 Unknown RAPPER mode 'rappermc/params.xml' requested.

This problem is easy to fix. There are really two rapper programs -- the rappermc/rapper shell script and rappermc/distrib/rapper_ARCH_CPU binaries. The rapper binaries need the first argument to be the path to the params.xml file, while rappermc/rapper calls the binaries with the params.xml file. So the command:

rappermc/rapper mode

is really the same as:

rappermc/distrib/rapper_ARCH_CPU rappermc/distrib/params.xml mode

You are calling rappermc/rapper rappermc/distrib/params.xml help, which is equivalent to:

rappermc/distrib/rapper_ARCH_CPU rappermc/distrib/params.xml mode rappermc/distrib/params.xml

All you need to do is stop providing the params.xml argument to the rapper script, and make sure the script is calling the rapper binary with the right params.xml file.

© 2001-2006 The RAPPER Team 
[Powered by FreeBSD]