On fine-tuning Boltz-2 for protein-protein affinity prediction

TL;DR

This paper evaluates Boltz-2, a structure-based model, for protein-protein affinity prediction, revealing its limitations compared to sequence-based methods and the benefits of combining different embedding types.

Contribution

It adapts Boltz-2 for protein-protein affinity prediction and demonstrates the complementary nature of structure- and sequence-based embeddings.

Findings

Boltz-2 underperforms compared to sequence-based models.

Combining embeddings improves prediction, especially for weaker models.

Structural data biases affect model performance.

Abstract

Accurate prediction of protein-protein binding affinity is vital for understanding molecular interactions and designing therapeutics. We adapt Boltz-2, a state-of-the-art structure-based protein-ligand affinity predictor, for protein-protein affinity regression and evaluate it on two datasets, TCR3d and PPB-affinity. Despite high structural accuracy, Boltz-2-PPI underperforms relative to sequence-based alternatives in both small- and larger-scale data regimes. Combining embeddings from Boltz-2-PPI with sequence-based embeddings yields complementary improvements, particularly for weaker sequence models, suggesting different signals are learned by sequence- and structure-based models. Our results echo known biases associated with training with structural data and suggest that current structure-based representations are not primed for performant affinity prediction.