SAFE (or Spatial Analysis of Functional Enrichment) is an automated network annotation algorithm. Given a biological network and a set of functional groups or quantitative features of interest, SAFE performs local enrichment analysis to determine which regions of the network are over-represented for each group or feature. SAFE visualizes the network and maps the detected enrichments onto the network.
SAFE was originally implemented in MATLAB and stored at https://bitbucket.org/abarysh/safe/. However, as of early 2017, the MATLAB implementation is only maintained for legacy purposes. New work related to SAFE is moving to Python and this repository.
WARNING. This package is still in development. Please use caution.
SAFE requires Python 3 and a set of packages listed in extras/requirements.txt. We recommend setting up a virtual environment and installing all the required packages via pip:
cd safepy/
virtualenv -p python3 safepy_env
source safepy_env/bin/activate
pip install git+https://github.com/baryshnikova-lab/safepy.git
After the installation is complete, it is useful to run a "hello world" SAFE analysis using the Jupyter notebook at examples/Example_1_GI_network_doxorubicin.ipynb.
To do so, from within the safepy_env environment install a new ipython kernel:
ipython kernel install --user --name=safepy_env
Then start jupyter, open examples/Example_1_GI_network_doxorubicin.ipynb and select the safepy_env kernel.
jupyter-notebook
To import safepy package in the jupyter notebook:
from safepy import safe
To run the examples, several common datasets will be required (e.g., the genetic interaction similarity network from Costanzo et al., 2016). These datasets are stored separately at https://github.com/baryshnikova-lab/safe-data (to avoid duplication with other SAFE-related repositories and packages). We recommend cloning the safe-data repository and storing it locally. In addition, it is necessary to edit the SAFE settings file (at safepy/safe_default.ini) with the path to the safe-data folder. By default, safe-data is expected to be located in the same parent folder as safepy:
vim safe_default.ini
...
[DEFAULT]
safe_data = ../safe-data/
...
It may also be useful to run a series of unit tests to verify that SAFE provides the correct outputs for default inputs. Tests are progressively being written and added to the repository. To run all the existing tests (from the safepy folder):
git package clone https://github.com/baryshnikova-lab/safepy.git
git clone https://github.com/baryshnikova-lab/safe-data.git
cd safepy/
pip install -e . ## install the package in developer mode
cd tests/
python -m unittest discover -v -s .
Please direct all questions/comments to Anastasia Baryshnikova (abaryshnikova@calicolabs.com).
The main repository for this code is at https://github.com/baryshnikova-lab/safepy. Please subscribe to the repository to receive live updates about new code releases and bug reports.
The manuscript describing SAFE and its applications is available at:
Baryshnikova, A. (2016). Systematic Functional Annotation and Visualization of Biological Networks. Cell Systems. http://doi.org/10.1016/j.cels.2016.04.014