GraPharm aims to learn new biological links between diseases, compounds, genes, pathways, biological process, etc. to unveil the hidden biological relations, accelerating drug discovery.

Prerequisites

To run the code in this respository, make sure you have installed Miniconda/Anaconda. Then, follow these steps to install the required packages:

• Download this repo: git clone https://github.com/GraPharm-ML/grapharm
• Change dir to grapharm folder.
• Install required packages inside virtual environment: bash install.sh
• Later for new updates in the package: pip install -e .

Streamlit application

To run application: streamlit run GraPharm_streamlit.py

Notebooks

To add conda virlenv to Jupyter kernel, first activate the grapharm env then type: python -m ipykernel install --user --name grapharm

• 01_hetionet_analysis.ipynb: Data analysis for Hetionet
• 02_ultra.ipynb: Analysis of the model result

References

• ULTRA: Towars Foundation Models for Knowledge Graph Reasoning
• Hetionet - An integrative network of biomedical knowledge

Acknowledgement

• Thank Totoro for very useful discussions about state-of-the-art methods for Graph Reasoning