Joint Binding Deconvolution

Downloading and Compiling JBD

You will need a recent version of Matlab to compile JBD. While you can run JBD directly in Matlab, you'll want a compiled version for most purposes.

Download the JBD code
Download the TAMO code
Untar the JBD code. You will need to add some subdirectories in the jbd directory to your matlab path. See the example startup.m as an example
Compile JBD. From the jbd/agilent directory, run make jbd.


If you're having trouble with the compiler, make sure you have a recent
version.  We've used version 2006a with good results, but various 6.x
versions don't work.  Also, you may need to modify the jbd/agilent/mbuildopts.sh 
file to make mcc use the system versions of various libraries rather than the versions that 
come with matlab.  

Running JBD

Put the input files in a directory
Run ep_cmdline.pl to generate scripts
that will run JBD.  This step is particularly useful if you plan to run
JBD on a cluster.
After running the scripts that ep_cmdline.pl produces, you should
have a set of jbd output files.
jbd2bindingevents.pl will process JBD output files to
give a summary of discrete binding events.
Determining significance of binding events (aka "where are my p-values")
jbd2motifinputs.pl will process JBD output files for
use with motif discovery.


If you're trying to run JBD from the command line without using the scripts that ep_cmdline.pl creates, you'll probably need to set your LD_LIBRARY_PATH to include 

the jbd/agilent directory (the one in which you compiled JBD)
bin/glnx86 or bin/glnxa64 in your matlab distribution


Running TAMO
The TAMO package 
takes the positional priors and input sequences from JBD to run motif
discovery.  See the Fraenkel Lab
pages for instructions on setting it up.


Instructions on
running TAMO
Description
of the output format



JBD uses the Lightspeed
MATLAB library.


Contact Us

Computational Genomics Research
Group, the Fraenkel Lab, and
the Young Lab.