Installing TADBit on GNU/Linux

Contents

• Installing TADBit on GNU/Linux
  • Dependencies
    • Python libraries
    • IMP - 3D modeling
    • MCL - clustering
    • Chimera - visualization
    • LiftOver
  • TADBit

Note: at the moment the installation has been tested only under Ubuntu-linux and MacOS (tested under OSX 10.9 with MacPorts www.macports.org).

TADBit requires python >= 2.6 as well as several dependencies that are listed below.

Dependencies

Python libraries

Required:

apt-get install python-scipy
apt-get install python-numpy

Optional packages (but highly recommended):

apt-get install python-matplotlib

Note

Alternative install, you can install python-setuptools and use easy_install to get these packages (e.g. “easy_install scipy”).

IMP - 3D modeling

The three-dimensional (3D) structure of a given genomic region is done via the IMP package. TADBit has been tested with IMP version 2.0.1 (recommended IMP version to be installed). The installation procedure reported in this manual describes how to install IMP on Ubuntu machines (installation tested on Ubuntu versions 12.04 and 13.04). However, we do not provide support for installing IMP; installation instructions for IMP can be found on the IMP website (http://salilab.org/imp/nightly/doc/html/).

Install the required libraries:

sudo apt-get install cmake
sudo apt-get install libboost1.49-all-dev
sudo apt-get install libhdf5-dev
sudo apt-get install swig
sudo apt-get install libcgal-dev
sudo apt-get install python-dev

Download the IMP tarball file from http://salilab.org/imp/ and uncompress it:

wget http://salilab.org/imp/get.php?pkg=2.0.1/download/imp-2.0.1.tar.gz -O imp-2.0.1.tar.gz
tar xzvf imp-2.0.1.tar.gz

Move into the IMP directory and compile the code (Note: the -j option stands for the number of CPUs you want to assign to the compiler; the higher the faster).

cd imp-2.0.1
cmake . -DCMAKE_BUILD_TYPE=Release -DIMP_MAX_CHECKS=NONE -DIMP_MAX_LOG=SILENT
make -j4

Once the compilation has finished, open the file setup_environment.sh in your IMP directory and copy the first lines into your ~/.bashrc file (if this file in not present in your home directory, create it). These lines should look like:

LD_LIBRARY_PATH="/SOMETHING/imp-2.0.1/lib:/SOMETHING/imp-2.0.1/lib:/SOMETHING/imp-2.0.1/src/dependency/RMF/:$LD_LIBRARY_PATH"
export LD_LIBRARY_PATH

PYTHONPATH="/SOMETHING/imp-2.0.1/lib:/SOMETHING/imp-2.0.1/lib:/SOMETHING/imp-2.0.1/src/dependency/RMF/:$PYTHONPATH"
export PYTHONPATH

Note

Important note Do not copy the lines above, copy them from setup_environment.sh, where SOMETHING is replaced by your real path to IMP.

MCL - clustering

MCL is the program used for clustering the 3D models generated by IMP. It can be downloaded from http://micans.org/mcl/; on Debian/Ubuntu machines it can be automatically installed with:

sudo apt-get install mcl

Note: if the MCL executable is not found by TADBit, an alternative clustering method will be used. Nevertheless we strongly recommend to use MCL.

Chimera - visualization

Chimera is a program used for visualization and analysis of molecular structures. It is used in TADBit to visualize the generated 3D models. Chimera is available at: http://www.cgl.ucsf.edu/chimera/

This software is only needed for the visualization of 3D models from inside TADBit.

LiftOver

TADBit provides a wrapper for the LiftOver tool [Fujita2011] (download it from: http://hgdownload.cse.ucsc.edu/admin/exe/ ).

This can be used to ease the conversion of genomic TAD coordinates (e.g.: to align human TADs with mouse TADs).

Note

A ‘chain’ file may also be downloaded. For example, to convert coordinates to hg19, the chain file needed may be found at: http://hgdownload.cse.ucsc.edu/goldenPath/hg19/liftOver/

TADBit

Once all the needed library/software have been installed, TADBit can be downloaded, unpacked and installed as:

wget https://github.com/3DGenomes/tadbit/archive/master.zip -O tadbit.zip
unzip tadbit.zip
cd tadbit-master/src
sudo python setup.py install

Note

IMP not found problem

If you are under debian/Ubuntu machines, and you have followed the quick tutorial proposed here to install IMP, you may encounter some warning message stating that IMP is not installed. This is because root user is not using the PYTHONPATH declared in your bashrc file. You can perfectly ignore the warning, or just proceed like this:

sudo PYTHONPATH=$PYTHONPATH python setup.py install

Finally, run the test script to check that the installation completed successfully. To do so, move to the test directory and run:

cd ../test
python test_all.py

[Fujita2011]

Fujita, P. A., Rhead, B., Zweig, A. S., Hinrichs, A. S., Karolchik, D., Cline, M. S., Goldman, M., et al. (2011). The UCSC Genome Browser database: update 2011. Nucleic Acids Research, 39(Database issue), D876-82. doi:10.1093/nar/gkq963