Bacterial protein function prediction via multimodal deep learning - Datasets & Materials

This record provides the resources for “Bacterial protein function prediction via multimodal deep learning”: GO label matrices (matrix_labels.zip), precomputed inputs/embeddings (features.zip, all_proteins.pkl), protein sequences (protein_sequences.zip), protein structures (structure.zip), and cross-validation splits (splits_foldseek.zip, structure-similarity/Foldseek, per species). Code: https://github.com/BorgwardtLab/DeepEST Data sources: Protein structures: AlphaFold Protein Structure Database (Varadi et al., 2021). GO annotations / protein metadata: UniProt (The UniProt Consortium, 2022). Gene expression: PATHOgenex RNA atlas / stress-condition expression dataset (Avican et al., 2021). References: Varadi, Mihaly, Anyango, Stephen, Deshpande, Mandar, et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Research50(D1): D439–D444 (2022). DOI: 10.1093/nar/gkab1061. The UniProt Consortium. UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Research 51(D1): D523–D531 (2023). DOI: 10.1093/nar/gkac1052. Avican, Kemal, Aldahdooh, Jehad, Togninalli, Matteo, et al. RNA atlas of human bacterial pathogens uncovers stress dynamics linked to infection. Nature Communications 12: 3282 (2021). DOI: 10.1038/s41467-021-23588-w.