Investigating the ability of deep learning-based structure prediction to extrapolate and/or enrich the set of antibody CDR canonical forms

Journal article

Summary: In this article, we introduce ABodyBuilder3, an improved and scalable antibody structure prediction model based on ABodyBuilder2. 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. Availability and implementation: The software package is available at https://github.com/Exscientia/ABodyBuilder3 with model weights and data at https://zenodo.org/records/11354577

Authors: Greenshields-Watson, A; Abanades, B; Deane, CM

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Frontiers Media
• Journal: Frontiers in Immunology
• Volume: 15
• Pages: 1352703-1352703
• Article number: 1352703
• Publication date: 2024-02-28
• DOI: 10.3389/fimmu.2024.1352703
• EISSN: 1664-3224
• ISSN: 1664-3224
• Language: English
• Keywords: Pattern recognition (psychology); Set (abstract data type); Genetics; Deep learning; Sequence (biology); Artificial intelligence; Computational biology; Training set; Algorithm; Biology; Computer science