Installation

Requirements

genipe requires a standard Python 3.7 (or latest) installation with the following modules:

  • numpy

  • jinja2

  • pandas

  • setuptools

The following modules are optional (required for statistical analysis and data management):

  • matplotlib

  • scipy

  • patsy

  • statsmodels

  • lifelines

  • biopython

  • pyfaidx

  • drmaa

  • pyplink

Note

Note that statsmodels (specifically MixedLM analysis) version 0.6 is not compatible with numpy version 1.12 and latest.

Note

Only Python version 3.7 and higher are supported.

Installing in a virtual environment

Using python’s pyvenv

The following commands should successfully create a virtual environment and activate it, as long as Python 3.7 (or latest) was previously installed on the machine.

# Creating the environment
pyvenv $HOME/genipe_pyvenv

# Activating the environment
source $HOME/genipe_pyvenv/bin/activate

Then, install the genipe package (and all its dependencies) using the following command:

pip install genipe

If required, the optional dependencies can be installed using the following command:

pip install scipy
pip install patsy
pip install statsmodels
pip install lifelines
pip install biopython
pip install pyfaidx
pip install matplotlib
pip install drmaa
pip install pyplink

Using Miniconda

The following commands should successfully download and install Miniconda, create and activate a new python virtual environment. This installation method doesn’t require a previous Python 3 installation.

# Installing Miniconda
wget http://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh
bash miniconda.sh -b -p $HOME/miniconda

# Creating the environment
$HOME/miniconda/bin/conda create -q -n genipe_pyvenv python=3

# Activating the environment
source $HOME/miniconda/bin/activate genipe_pyvenv

Then, install the genipe package (and all its dependencies) using the following command:

conda install genipe -c http://statgen.org/wp-content/uploads/Softwares/genipe

Note

It is possible to add the channel to conda’s configuration (so that you won’t need to use the -c option for installing or updating). To do so, perform the following command:

conda config --add channels http://statgen.org/wp-content/uploads/Softwares/genipe

Once this command is executed, you can always ommit -c http://statgen.org/... in the conda commands (for installing or updating).

If required, the optional dependencies can be installed using the following commands:

conda install -y scipy
conda install -y patsy
conda install -y statsmodels
conda install -y biopython
conda install -y matplotlib
conda install -y drmaa
pip install --no-deps pyfaidx
pip install --no-deps lifelines

Virtual environment activation

Before using the genipe module for any analysis, the Python virtual environment needs to be activated. Depending of your installation type (i.e pyvenv or miniconda), the activation process will differ.

Pyvenv

If the module was installed into a pyvenv environment, perform the following command to activate it.

source $HOME/genipe_pyvenv/bin/activate

Miniconda

If the module was installed into a Miniconda environment, perform the following command to activate it.

source $HOME/miniconda/bin/activate genipe_pyvenv

Testing the installation

The genipe module has been tested with the most recent versions of the requirements. To test the installation, make sure that the virtual environment is activated. Then, launch Python and use the following python commands:

>>> import genipe
>>> genipe.test()

Updating the package

If there is a new genipe release, perform one of the following commands (depending of the installation method). Don’t forget to first activate the python virtual environment.

Using python’s pyvenv

pip install -U genipe

Using Miniconda

conda update genipe -c http://statgen.org/wp-content/uploads/Softwares/genipe

Note

If you have configured conda to use the genipe channel (see the note above), the following command can be executed to update the package:

conda update genipe