GNN encodings of bovine metabolities towards anti-methanogenicity

This project focuses on analyzing and visualizing the complex molecular structures present in milk and bovine metabolites. We use Graph Neural Networks (GNNs) to generate molecular embeddings from SMILES string and t-SNE for dimensionality reduction and visualization.

We utilize two metabolite databases: The Milk Composition Database (MCDB) - 2,360 entries The Bovine Metabolome Database (BMDB) - 51,682 entries Sample - 16 selected MCR molecules

Note: The MCDB database is included in the metabolites folder of this repository. Due to its large size, the BMDB is not incorporated directly into this repository. The databses are publicly available here: https://metabolomicscentre.ca/software-databases/databases/. Bovine metabolome database can also be accessed from: https://iastate.box.com/s/8xz7ipb6rxqdfgfpcurxyq0ynd5l9g29