This repository contains Python scripts for "deepNF: Deep network fusion for protein function prediction" by V. Gligorijevic, M. Barot and R. Bonneau.
@article {Gligorijevic2017,
author = {Gligorijevi{\'c}, Vladimir and Barot, Meet and Bonneau, Richard},
title = {deepNF: Deep network fusion for protein function prediction},
year = {2018},
doi = {10.1093/bioinformatics/bty440},
pages = {bty440},
publisher = {Oxford},
URL = {http://dx.doi.org/10.1093/bioinformatics/bty440},
journal = {Bioinformatics}
}
To run deepNF run the following command from the project directory:
python main.py
To see the list of options:
python main.py --help
To compute network emgeddings only use net_embedding.py script. Input file format: edgelist (i, j, w_ij)
For a single network:
python net_embedding.py --model_type ae --nets example_net_1.txt
For multiple networks:
python net_embedding.py --model_type mda --nets example_net_1.txt example_net_2.txt
deepNF is tested to work under Python 3.6.
The required dependencies for deepNF are Keras, TensorFlow, Numpy, NetworkX and scikit-learn.
Data (PPMI matrices for human and yeast STRING networks as well as protein annotations) used for producing figures in the paper can be downloaded from:
https://users.flatironinstitute.org/vgligorijevic/public_www/deepNF_data/