Sequence-only antigen context improves modeling of antibody hypervariable regions
- 1University of California, Riverside
Abstract
The complementarity-determining regions (CDRs) of antibodies govern antigen specificity; consequently, CDR design should be performed conditional on the target antigen. Most published approaches depend on the availability of the antigen's three-dimensional (3D) structure, which is often difficult to obtain experimentally. In contrast, sequence-only models circumvent this requirement but are typically trained on antibodies without incorporating antigen context; exposure to antigen sequences alone does not necessarily entail learning the underlying interaction determinants. Here, we introduce LangAAI, a masked language model that conditions exclusively on antibody and antigen primary sequences and therefore does not require 3D structures, docking poses, or epitope annotations. To assess whether the model exhibits genuine antigen conditioning, we replace the cognate antigen with a non-binding decoy. Whereas antigen-agnostic models are largely insensitive to this substitution, LangAAI shows a marked reduction in sequence recovery that is specifically concentrated in the hypervariable CDR3 loop. Across benchmarks, LangAAI improves amino-acid recovery relative to existing antibody language models, matches the performance of structure-based methods when evaluated under their respective protocols, and ranks experimentally measured variants by affinity comparably to inverse-folding models that use backbone coordinates. Collectively, these results support LangAAI as a structure-independent approach for antigen-aware antibody optimization, enabling prioritization of candidate variants for targets lacking solved antigen-antibody complexes.
Related
- Antibody-Antigen Affinity Prediction with Chain-Aware Protein Language Modeling
- Identification of a non-canonical G3BP-binding sequence in a Mayaro virus nsP3 hypervariable domain
- Co-administration of Staphylococcus aures Fibronectin Binding Protein A N1-3 region and Lipoteichoic acid antigen mixture enhance antibody response in BALB/c mice
- CD4+ T cells are required for tumor antigen-expressing oHSV glioma therapeutic effect and antigen-specific antibody production
- CHAIR an antigen-resolved antibody-reactome resource for healthy adult immune aging
BibTeX
@misc{ashraf2026sequenceonly,
title = {{Sequence-only antigen context improves modeling of antibody hypervariable regions}},
author = {F. B. Ashraf and S. Lonardi},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.19.760314},
url = {https://www.biorxiv.org/content/10.64898/2026.10.19.760314v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Sequence-only antigen context improves modeling of antibody hypervariable regions
AU - F. B. Ashraf
AU - S. Lonardi
PY - 2026
DA - 2026/10/19
PB - bioRxiv
DO - 10.64898/2026.10.19.760314
UR - https://www.biorxiv.org/content/10.64898/2026.10.19.760314v1
AB - The complementarity-determining regions (CDRs) of antibodies govern antigen specificity; consequently, CDR design should be performed conditional on the target antigen. Most published approaches depend on the availability of the antigen's three-dimensional (3D) structure, which is often difficult to obtain experimentally. In contrast, sequence-only models circumvent this requirement but are typically trained on antibodies without incorporating antigen context; exposure to antigen sequences alone does not necessarily entail learning the underlying interaction determinants. Here, we introduce LangAAI, a masked language model that conditions exclusively on antibody and antigen primary sequences and therefore does not require 3D structures, docking poses, or epitope annotations. To assess whether the model exhibits genuine antigen conditioning, we replace the cognate antigen with a non-binding decoy. Whereas antigen-agnostic models are largely insensitive to this substitution, LangAAI shows a marked reduction in sequence recovery that is specifically concentrated in the hypervariable CDR3 loop. Across benchmarks, LangAAI improves amino-acid recovery relative to existing antibody language models, matches the performance of structure-based methods when evaluated under their respective protocols, and ranks experimentally measured variants by affinity comparably to inverse-folding models that use backbone coordinates. Collectively, these results support LangAAI as a structure-independent approach for antigen-aware antibody optimization, enabling prioritization of candidate variants for targets lacking solved antigen-antibody complexes.
ER -