Decagon is a graph convolutional neural network (GCN) approach for multirelational link prediction in multimodal graphs.
It is a general approach for multirelational link prediction in any multimodal network. Decagon handles multimodal graphs with large numbers of edge types. It works on simple and multimodal graphs.
Please check the project page for more details, including the preprocessed datasets for modeling drug combinations.
Decagon has been used to solve problems in computational pharmacology, specifically to model polypharmacy side effects of drug pairs (i.e., drug combinations).
Using Decagon, we construct a multimodal graph of protein-protein interactions, drug-protein target interactions, and the polypharmacy side effects, which are represented as drug-drug interactions, where each side effect is an edge of a different type. Decagon then predicts the exact side effect, if any, through which a given drug combination manifests clinically.
Decagon species a graph convolutional neural network architecture based on this multimodal graph to model and predict polypharmacy side effects.
The setup for this problem using a dummy dataset is outlined in:
main.py
All preprocessed datasets used for polypharmacy side effect prediction are available for download from project page.
If you find Decagon useful for your research, please consider citing this paper:
@article{Zitnik2018,
title = {Modeling polypharmacy side effects with graph convolutional networks.},
author = {Zitnik, Marinka and Agrawal, Monica and Leskovec, Jure},
journal = {Bioinformatics},
volume = {34},
number = {13},
pages = {457–466},
year = {2018}
}
Please send any questions you might have about the code and/or the algorithm to [email protected].
Decagon is tested to work under Python 2.7.
The required dependencies for Decagon are NumPy >= 1.13, NetworkX >= 2.0, and TensorFlow >= 1.1.
Decagon is licensed under the MIT License.