Generalise or Memorise? Benchmarking Ligand-Conditioned Protein Generation from Sequence-Only Data

Proteins can bind small molecules with high specificity. However, designing proteins that bind userdefined ligands remains a challenge, typically relying on structural information and costly experimental iteration. While protein language models (pLMs) have shown promise for unconditional generation and conditioning on coarse functional labels, instance-level conditioning on a specific ligand has not been evaluated using purely textual inputs. Here we frame small-molecule protein binder design as a sequence-to-sequence translation problem and train ligand-conditioned pLMs that map molecular strings to candidate binder sequences. We curate large-scale ligand-protein datasets (>17M ligand-protein pairs) covering different data regimes and train a suite of models, spanning 16 to 700M parameters. Results reveal a consistent trade-off driven by supervision ambiguity: when each ligand is paired with few proteins, models generate near-neighbour, foldable sequences; when each ligand is paired with many proteins, generations are more diverse but less consistently foldable. Our study exposes how annotation diversity and sampling choices elicit this behaviour and how it changes with the data distribution. These insights highlight dataset redundancy and incompleteness as key bottlenecks for sequence-only binder design. We release the curated datasets, trained models, and evaluation tools to support future work on ligand-conditioned protein generation.