Evaluating deep learning based structure prediction methods on antibody–antigen complexes
Bioinformatics · 07.04.2026 · 42 (4)
DOI: 10.1093/bioinformatics/btag136 ↗Kurz gefasst
Fachartikel in der Zeitschrift Bioinformatics, online erschienen am 21.03.2026. Verfasst von Samuel Fromm und 2 weiteren Autor:innen. Beteiligte Einrichtung: Science for Life Laboratory and Department of Biochemistry and Biophysics, Stockholm University, Solna 171 21, Sweden. Hauptthemen laut MEDLINE: Antigen-Antibody Complex, Deep Learning, Computational Biology. Die Arbeit wurde bisher 23-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Vetenskapsrådet, Grant no. 2021-03979, and the Knut and Alice Wallenberg Foundation.
Abstract (Originalsprache)
Motivation AlphaFold2 significantly improved the prediction of protein complex structures. However, its accuracy is lower for interactions without coevolutionary signals, such as host–pathogen and antibody–antigen interactions. Two strategies have been developed to address this limitation: massive sampling and replacing the evoformer with the pairformer, which does not rely on coevolution, as introduced in AlphaFold3, thereby enabling more structural reasoning by the network. Results In this study, we benchmark structure prediction methods on unseen antibody–antigen complexes. We found that increased sampling improves the chances of generating a correct protein model, roughly in a log-linear manner. However, the internal quality estimates by AlphaFold often cannot identify the best predicted structures for each target, resulting in a significant loss of performance for the top-ranked protein model compared with the best model. For all methods, a significant challenge remains the identification of the best model. We also show that AlphaFold3 outperforms AlphaFold2, Boltz-1, and Chai-1. Furthermore, AlphaFold3 performance declines significantly for complexes that lack structural similarity to the training set, indicating that it has to some extent learned to detect remote structural similarities. Availability and implementation All code is available from github.com/samuelfromm/abag-benchmark-set/ and all data from DOI: 10.5281/zenodo.17978681. The latter repository also contains the code.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Science for Life Laboratory and Department of Biochemistry and Biophysics, Stockholm University, Solna 171 21, Sweden.
ORCID-Profile: Arne Elofsson
Förderung
- Vetenskapsrådet, Grant no. 2021-03979, and the Knut and Alice Wallenberg Foundation
Zitieren
APA
Fromm, S., Ludaic, M., & Elofsson, A. (2026). Evaluating deep learning based structure prediction methods on antibody–antigen complexes. Bioinformatics, 42(4). https://doi.org/10.1093/bioinformatics/btag136
Vancouver
Fromm S, Ludaic M, Elofsson A. Evaluating deep learning based structure prediction methods on antibody–antigen complexes. Bioinformatics. 2026;42(4). doi:10.1093/bioinformatics/btag136
BibTeX
@article{Fromm2026,
author = {Fromm, Samuel and Ludaic, Marko and Elofsson, Arne},
title = {Evaluating deep learning based structure prediction methods on antibody–antigen complexes},
journal = {Bioinformatics},
year = {2026},
volume = {42},
number = {4},
doi = {10.1093/bioinformatics/btag136}
}
EUVIMED