--- license: apache-2.0 --- Substrate-Dataset as defined in *"Generalise or Memorise? Benchmarking Ligand-Conditioned Protein Generation from Sequence-Only Data"* Our dataset is curated from the following sources: - **Rhea** (Bansal *et al.*, 2022) - **BKMS** (Lang *et al.*, 2011) - **UniProt** (Bateman *et al.*, 2025) - **BRENDA** (Chang *et al.*, 2021) ## Citation Anonymised for double-blind review