AbRefine: Framework-Conditioned Refinement for Structure-Guided Antibody CDR Design

Abstract

Structure-guided antibody CDR design constrains residue selection using backbone geometry, but structural evidence alone may leave multiple amino acids plausible. We investigate whether the fixed antibody framework provides complementary sequence information for resolving this ambiguity. We introduce AbRefine, an information-aware fusion framework that combines backbone-conditioned structural predictions with framework-conditioned sequence distributions derived from a pretrained antibody language model with all designed CDRs masked. Its Information-Aware Fusion (IAF) module learns residue-specific fusion weights from the uncertainty, disagreement, and information content of the two predictive distributions, while keeping both pretrained experts frozen. Controlled interventions show that the signal is target specific: matched frameworks outperform shuffled frameworks by 7.37 percentage points, and the largest gain occurs at positions with high structural uncertainty and informative framework evidence (+5.77$ points). Compared with static fusion, IAF further improves AbMPNN recovery by +1.53 points (95\% CI +0.93$ to +2.14$) and significantly reduces native negative log-likelihood for both AbMPNN and AntiFold, without detectable degradation in fold-back structural compatibility. These results show that framework context provides information beyond backbone geometry and that learning its residue-specific utility enables effective adaptive fusion for antibody CDR design.

Pocket contrastive learning and conformation fusion for robust structure-conditioned 3D molecular generation under receptor conformational variation.
Huan Liu, Yijie Ding, Xiaoyi Guo +1
Bioinformatics (Oxford, England)1 Oct 2026Peer-reviewedCC BYFull textBioinformatics & computational biologyEurope PMCCrossref
Inverse FoldDir: Structure-conditioned Protein Sequence Design by Dirichlet Flow Matching
A. TARTICI, M. Stojkovic, A. Tian +3
bioRxiv6 Sep 2026PreprintCC BYFull textBioinformatics & computational biologybioRxiv
Flora-YOLO: A lightweight framework for real-time rose grading on edge devices via structural optimization and feature refinement
Wenwei Liu, Jinyu Xu, Guoao Wang +6
PLOS One21 Sep 2026Peer-reviewedCC BY 4.0Full textPlant biologyPLOS
X-SPEAR: An Extensible Framework for Spatial Panel Evaluation and Refinement
R. Patel, B. Duong, V. Patel +3
bioRxiv13 Oct 2026PreprintPublic domainFull textGenetics & genomicsbioRxiv
Relation-Aware Graph-Hypergraph Networks for Antibody Sequence-Structure Co-Design
J. Wei, S. Zhang, H. Duan +2
bioRxiv12 Oct 2026PreprintCC BYFull textBioinformatics & computational biologybioRxiv

Cite

BibTeX RIS

BibTeX

@misc{liu2026abrefine,
  title = {{AbRefine: Framework-Conditioned Refinement for Structure-Guided Antibody CDR Design}},
  author = {Q. Liu and X. Xu and Q. Cheng and L. Meng and C. Wu and D. Yang and X. Gao and X. Chen},
  howpublished = {bioRxiv},
  year = {2026},
  month = sep,
  doi = {10.64898/2026.09.27.754743},
  url = {https://www.biorxiv.org/content/10.64898/2026.09.27.754743v1},
  note = {CC BY-NC-ND},
}

RIS

TY  - UNPB
TI  - AbRefine: Framework-Conditioned Refinement for Structure-Guided Antibody CDR Design
AU  - Q. Liu
AU  - X. Xu
AU  - Q. Cheng
AU  - L. Meng
AU  - C. Wu
AU  - D. Yang
AU  - X. Gao
AU  - X. Chen
PY  - 2026
DA  - 2026/09/27
PB  - bioRxiv
DO  - 10.64898/2026.09.27.754743
UR  - https://www.biorxiv.org/content/10.64898/2026.09.27.754743v1
AB  - Structure-guided antibody CDR design constrains residue selection using backbone geometry, but structural evidence alone may leave multiple amino acids plausible. We investigate whether the fixed antibody framework provides complementary sequence information for resolving this ambiguity. We introduce AbRefine, an information-aware fusion framework that combines backbone-conditioned structural predictions with framework-conditioned sequence distributions derived from a pretrained antibody language model with all designed CDRs masked. Its Information-Aware Fusion (IAF) module learns residue-specific fusion weights from the uncertainty, disagreement, and information content of the two predictive distributions, while keeping both pretrained experts frozen. Controlled interventions show that the signal is target specific: matched frameworks outperform shuffled frameworks by 7.37 percentage points, and the largest gain occurs at positions with high structural uncertainty and informative framework evidence (+5.77$ points). Compared with static fusion, IAF further improves AbMPNN recovery by +1.53 points (95\% CI +0.93$ to +2.14$) and significantly reduces native negative log-likelihood for both AbMPNN and AntiFold, without detectable degradation in fold-back structural compatibility. These results show that framework context provides information beyond backbone geometry and that learning its residue-specific utility enables effective adaptive fusion for antibody CDR design.
ER  -

Abstract

Related

Cite