How do I install InVEx?
InVEx requires 64-bit Python 2.6 (or 2.7) with the following packages installed: NumPy, SciPy, Cython, and PyFasta. The example installation below was performed on a Mac running OS X 10.6.8 Snow Leopard. Ideally, the installation would be simplified by installing a free Python distribution that includes NumPy and SciPy (such as the EPD - Enthought Python Distribution). However, as of this writing, this distribution is only available in 32-bit mode for OS X 10.6.8. Therefore, downloading python and additional packages and installing from source code is required as described below. Ideally, this process is simplified on Linux, Windows, or the latest Mac OS X 10.7 Lion. However, instructions for installation on these systems is not provided. (Update: Windows 7 Installation Instructions provided by Shyam S Jayaraman MD at LABioMed Research Institute.)
The post-processing and report generation code was written and tested using R-2.15.0. Finally, the report generation R script requires the Nozzle library. An R installation description is provided below.
This section describes installation on a Mac running OS X 10.6.8 Snow Leopard.
XCode is required for easily building Python and Cython from source. It can be obtained by opening the Mac App Store and searching for XCode. Or you can navigate to the Mac Dev Center, login, and click on View in Mac App Store next to Xcode 4.3.3 for Lion. Although this mentions Lion, this download should install properly on Snow Leopard.
If you already have Python 2.6 installed on your machine, you can check if it is 32 or 64 bit by running the following command: python -c "import sys; print sys.maxint;". If 9223372036854775807 is displayed, you have 64 bit installed and you can move on to installing NumPy. If 2147483647 is displayed, you have 32 bit installed and should keep reading . Start by downloading the 2.6.6 source code. Do not download the Mac Installer disk image - it will install 32-bit Python. The instructions below were taken from this site: Web Amused.
First, assuming you are using bash, add the following two lines to your .bash_profile:
Then, untar the python source (tar xvzf Python-2.6.6.tgz) and cd into the Python-2.6.6 directory. Edit /etc/paths and move /usr/local/bin to the top of the file. Then run the following:
./configure --disable-framework --disable-toolbox-glue \
Next, run make followed by sudo make install. Finally, source your .bash_profile that you modified above in your home directory. If Python installed correctly, you should see something similar to this when you run python:
Python 2.6.6 (r266:84292, Jun 4 2012, 13:44:44)
Run python -c "import sys; print sys.maxint;" again to ensure you get 9223372036854775807.
Download the NumPy v1.6.2 source code. Untar the source: tar xvzf numpy-1.6.2.tar.gz and cd into numpy-1.6.2. Then run sudo python setup.py install. Run python and import numpy to ensure the installation succeeded.
SciPy installation requires a FORTRAN compiler. Download the GFortran v4.6.2 disk image. Double-click on the file and follow the installation instructions.
Download the SciPy v0.10.1 source code. Untar the source: tar xvzf scipy-0.10.1.tar.gz and cd into scipy-0.10.1. Then run sudo python setup.py install. Run python and import scipy to ensure the installation succeeded.
Download PyFasta v0.4.5. Untar the source: tar xvzf pyfasta0.4.5.tar.gz and cd into pyfasta-0.4.5. Then run sudo python setup.py install. Run python and import pyfasta to ensure the installation succeeded.
Download Cython v0.16. Untar the source: tar xvzf Cython-0.16.tar.gz and cd into Cython-0.16. Then run sudo python setup.py install. Run python and import Cython to ensure the installation succeeded.
Download R v2.15.0. Double-click on the obtained R-2.15.0.pkg file and follow the instructions to install. Then, download tcltk-18.104.22.168-x11.dmg. Double-click on the obtained tcltk.pkg file and follow the instructions to install this library.
Start an R session and type the following two lines at the command prompt to install the qvalue package.
Download Nozzle 0.3-0. Untarring the source: tar xvzf Nozzle.R1_0.3-0.tar.gz should result in a Nozzle.R1 directory. In the directory that contains Nozzle.R1, run the following command to install.
R CMD install Nozzle.R1
InVEx Source Installation
If you haven't followed the instructions for obtaining the source code yet, please Download first and then continue. Otherwise, unpackage the InVEx-#.#.#.tar.gz file and cd into the InVEx-#.#.# directory. Then run python setup.py build_ext --inplace. This uses Cython to build the c_permute.c and c_permute.so files from c_permute.pyx. You should now be ready to start running InVEx.