 |
 If
you work at Scripps, you can use the latest version of AutoDock
that lives under the '/mgl' directory.
You do not need to sign a license agreement. The latest version is developed
on an SGI, so you should try to run AutoDock on an SGI running IRIX64
if you want the most up-to-date version. AutoDock is also available
on the SGI Origin Compute Cluster,
accessible via
"atlas". These machines currently
run the PBS batch
system, and are a good resource for heavy number crunching, like AutoDock
dockings. If you need to run on another platform, there are executables
but they may not be completely up-to-date: I can always re-compile the
latest source if you need it on another platform.
If you need help, or have a request for a new feature, let me know and I will make sure that your question is answered as soon as possible. Garrett M. Morris
(1) The
machine you wish to run AutoDock should be able to automount '/mgl'
and its sub-directories. Check if you can do this:
% cd /mglIf you can, then you go to step (2). If not., you will have to install a local copy on your machine. (2) If you want to know what type of platform your machine is (i.e. hardware and operating system versions, etc.), type: % uname -a(3) You should copy the '.autodocksetup' file to your home directory: % cp /home/garrett/.autodocksetup $HOME(4) Next, in your '.login' file, add these lines: setenv ARCHOSV sgi4DIRIX646 #if using SGI IRIX64.This should set up everything you need to use AutoDock and its related programs ($AUTODOCK_BIN) and scripts ($AUTODOCK_UTI). If you are running on the Power Challenge via "atlas" then add this to your '.login' file (in the SGI section if you have one): set path = ($path /usr/people/applications/autodock $AUTODOCK_UTI)(5) Now you are ready to get started. Download a PostScript copy of the AutoDock User Guide and print it out, or read it online. If you do not have time to read documentation, jump right to the `Getting Started' section of the User Guide. Note that on the PowerChallenge cluster, you can launch several AutoDock jobs at once using the "submit.py" script, and while the dockings are running or when they're finished, you can cluster the results using "recluster.py" into one large histogram. If you already have a DPF with the name "stem.dpf", and you want to run on ncpus CPUs, where ncpus is an integer, say 10 or 20, then type this: % submit.py stem ncpus To re-cluster the dockings, use ADT or on the command-line on 'atlas', the usage is: % recluster.py stem num ['during'|'end'] [rmstol] For example: % recluster.py z3 20 during 3.5 So here, z3.dpf is the name of the DPF, 20 is the number of concurrent docking jobs you want re-clustered, 'during' means that results produced during the dockings will be used, and '3.5' is the RMSD tolerance in Angstroms for re-clustering. Note that 'recluster.py' is dependent on you using 'submit.py' to start the jobs, since it makes the same assumptions about file naming conventions. Happy docking! |
