The results can be reproduced by running this. I experimented with using a transformer encoder and decoder but they seemed to perform worse and that makes sense because they are not really meant to work on such network data. Better embeddings could be generated if the NetMF/Deepwalk matrix can be understood but at the moment, I dont have any ideas on how to improve the algorithm or the model.
This repository contains the scripts to run the ETNA method and corresponding analysis described in the Li et al. paper, Joint embedding of biological networks for cross-species functional alignment.
Joint embedding of biological networks for cross-species functional alignment. Li L, Dannenfelser R, Zhu Y, Hejduk N, Segarra S, Yao V. BioRxiv. Jan 2022. https://doi.org/10.1101/2022.01.17.476697
Model organisms are widely used to better understand the molecular causes of human disease. While sequence similarity greatly aids this transfer, sequence similarity does not imply functional similarity, and thus, several current approaches incorporate protein-protein interactions to help map findings between species. Existing transfer methods either formulate the alignment problem as a matching problem which pits network features against known orthology, or more recently, as a joint embedding problem. Here, we propose a novel state-of-the-art joint embedding solution: Embeddings to Network Alignment (ETNA). More specifically, ETNA generates individual network embeddings based on network topological structures and then uses a Natural Language Processing-inspired cross-training approach to align the two embeddings using sequence orthologs. The final embedding preserves both within and between species gene functional relationships, and we demonstrate that it captures both pairwise and group functional relevance. In addition, ETNA’s embeddings can be used to transfer genetic interactions across species and identify phenotypic alignments, laying the groundwork for potential opportunities for drug repurposing and translational studies.
ETNA's method is roughly divided into 3 main parts:
- training an autoencoder to embed PPI networks
- aligning the embeddings between species via ortholog anchors
- scoring gene pairs across organisms with cosine similarity in the embedding
These steps are implemented in src/algorithms/ETNA.py
. The demo
jupyter notebook (/src/demo.ipynb
) illustrates running ETNA to align
two PPI networks from S. cerevisiae and S. pombe.
This project uses conda to manage the required packages and setup a virtual environment. Once conda is installed on your machine get started by setting up the virtual environment.
conda env create -f env.yml
conda activate etna
We have created demo code for running and evaluating ETNA on two PPI networks. To run the demo start up a jupyter notebook with the following command:
jupyter lab --port=8888
Navigate in a browser to running notebook at http://localhost:8888
and open the src
folder to load and run demo.py
.