ABodyBuilder3: Improved and scalable antibody structure predictions

TL;DR

ABodyBuilder3 is a scalable antibody structure prediction model that achieves state-of-the-art accuracy in modeling CDR loops by leveraging language model embeddings and improved relaxation strategies.

Contribution

It introduces ABodyBuilder3, which enhances antibody structure prediction accuracy and scalability using language models and uncertainty estimation methods.

Findings

State-of-the-art accuracy in CDR loop modeling

Enhanced structure predictions through relaxation strategies

Incorporation of uncertainty estimation in predictions

Abstract

Accurate prediction of antibody structure is a central task in the design and development of monoclonal antibodies, notably to understand both their developability and their binding properties. In this article, we introduce ABodyBuilder3, an improved and scalable antibody structure prediction model based on ImmuneBuilder. We achieve a new state-of-the-art accuracy in the modelling of CDR loops by leveraging language model embeddings, and show how predicted structures can be further improved through careful relaxation strategies. Finally, we incorporate a predicted Local Distance Difference Test into the model output to allow for a more accurate estimation of uncertainties.