Using Graph Neural Networks for Site-of-Metabolism Prediction and its Applications to Ranking Promiscuous Enzymatic Products
This repository contains data and code for running the GNN-SOM models for site-of-metabolism prediction.
We recommend using conda for managing this environment. Our implementation requires the following software toolkits to be installed: * PyTorch * PyTorch Geometric * RDKit
The example code for making SOM predictions on a given enzyme-molecule pair is presented as a Jupyter notebook, which can be found in GNN-SOM.ipynb. This notebook makes use of various commonly-used functions provided in the gnn_som directory as well as the model state and configuration files in the data directory.
This project is licensed under the MIT license.