EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies

Xin Yu iD, Kostika Vangjeli, Anusha Prakash, Meha Chhaya, Samantha J Stanley, Noah Cohen, Lili Huang

Antibody Therapeutics · 19.07.2024 · 7 (3) · 199-208

DOI: 10.1093/abt/tbae012 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Antibody Therapeutics, online erschienen am 30.05.2024. Verfasst von Xin Yu und 6 weiteren Autor:innen. Beteiligte Einrichtung: Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 23 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: AbbVie Inc..

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Early assessment of antibody off-target binding is essential for mitigating developability risks such as fast clearance, reduced efficacy, toxicity, and immunogenicity. The baculovirus particle (BVP) binding assay has been widely utilized to evaluate polyreactivity of antibodies. As a complementary approach, computational prediction of polyreactivity is desirable for counter-screening antibodies from in silico discovery campaigns. However, there is a lack of such models.

Methods

Herein, we present the development of an ensemble of three deep learning models based on two pan-protein foundational protein language models (ESM2 and ProtT5) and an antibody-specific protein language model (PLM) (Antiberty). These models were trained in a transfer learning network to predict the outcomes in the BVP assay and the bovine serum albumin binding assay, which was developed as a complement to the BVP assay. The training was conducted on a large dataset of antibody sequences augmented with experimental conditions, which were collected through a highly efficient application system.

Results

The resulting models demonstrated robust performance on canonical mAbs (monospecific with heavy and light chain), bispecific Abs, and single-domain Fc (VHH-Fc). PLMs outperformed a model built using molecular descriptors calculated from AlphaFold 2 predicted structures. Embeddings from the antibody-specific and foundational PLMs resulted in similar performance.

Conclusion

To our knowledge, this represents the first application of PLMs to predict assay data on bispecifics and VHH-Fcs.

Schlagwörter

Stichwörter der Autor:innen

bispecific antibodydeep learningpolyreactivitypolyspecificityprotein language model

Autor:innen und Institutionen

  • Biotherapeutics and Genetic Medicine, AbbVie Bioresearch Center, Worcester, MA 01605, United States.

ORCID-Profile: Xin Yu

Förderung

  • AbbVie Inc.

Zitieren

APA

Yu, X., Vangjeli, K., Prakash, A., Chhaya, M., Stanley, S. J., Cohen, N., & Huang, L. (2024). Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies. Antibody Therapeutics, 7(3), 199-208. https://doi.org/10.1093/abt/tbae012

Vancouver

Yu X, Vangjeli K, Prakash A, Chhaya M, Stanley SJ, Cohen N, et al. Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies. Antibody Therapeutics. 2024;7(3):199-208. doi:10.1093/abt/tbae012

BibTeX
@article{Yu2024,
  author = {Yu, Xin and Vangjeli, Kostika and Prakash, Anusha and Chhaya, Meha and Stanley, Samantha J and Cohen, Noah and Huang, Lili},
  title = {Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies},
  journal = {Antibody Therapeutics},
  year = {2024},
  volume = {7},
  number = {3},
  pages = {199--208},
  doi = {10.1093/abt/tbae012}
}
Kostenlos starten →