Multimodal reasoning for broadly neutralizing antibody discovery from label-free human B cell repertoires across virus families

@misc{lou2026multimodal,
  title = {{Multimodal reasoning for broadly neutralizing antibody discovery from label-free human B cell repertoires across virus families}},
  author = {Hantao Lou and Jianqing Zheng and Can Yue and Meihan Zhang and Yuanchao Bao and Yu Chen and Mengting Huang and Yupeng Yang and Qianyu Pan and Nana Fu and Yansong Shi and Hongli Li and Yangyang Chai and Ruyi Chen and Wansheng Li and Zhu Liang and Rongmei Yao and Yuanhan Mo and Lei Wang and Chunmei Wang and Yun Quan and Qiong Zhang and Xiangxi Wang and Xuetao Cao},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.03160},
  eprint = {2610.03160},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.03160},
}

TY  - UNPB
TI  - Multimodal reasoning for broadly neutralizing antibody discovery from label-free human B cell repertoires across virus families
AU  - Hantao Lou
AU  - Jianqing Zheng
AU  - Can Yue
AU  - Meihan Zhang
AU  - Yuanchao Bao
AU  - Yu Chen
AU  - Mengting Huang
AU  - Yupeng Yang
AU  - Qianyu Pan
AU  - Nana Fu
AU  - Yansong Shi
AU  - Hongli Li
AU  - Yangyang Chai
AU  - Ruyi Chen
AU  - Wansheng Li
AU  - Zhu Liang
AU  - Rongmei Yao
AU  - Yuanhan Mo
AU  - Lei Wang
AU  - Chunmei Wang
AU  - Yun Quan
AU  - Qiong Zhang
AU  - Xiangxi Wang
AU  - Xuetao Cao
PY  - 2026
DA  - 2026/10/02
PB  - arXiv
DO  - 10.48550/arxiv.2610.03160
UR  - https://arxiv.org/abs/2610.03160
AB  - Discovering broadly neutralizing antibodies (bnAbs) from human natural immune repertoires remains a fundamental challenge in immunology, hindered by: the extreme rarity of bnAb, incomplete understanding of their cellular origins across pathogens, and the inability of existing computational tools to generalize across emerging viral threats. Here we present ImmuneAgent, a closed-loop AI system that integrates multimodal reasoning with continual meta-learning and wet-lab feedback to overcome these barriers. Applied to screen the natural BCR repertoires from vaccinated or infected cohorts, the system achieves a ~55% neutralization antibody discovery rate (60 of 110 cloned candidates) and a ~11% bnAb yield (12 of 110), substantially outperforming a state-of-the-art sequence-based neutralization predictor or cofolding models evaluated at the same cloning budget. Five ImmuneAgent-discovered antibodies conferred 100% in vivo protection against lethal influenza challenge, comparable to the clinical-stage therapeutic MEDI8852. The system recovered the cellular and structural determinants of bnAb activity and identified FCRL5+CD27+ atypical memory B cells as a conserved bnAb reservoir and hydrophobic interface enrichment as a cross-viral structural signature, which generalized to unseen antigens, discovering human metapneumovirus (hMPV) cross-neutralizing and human papillomavirus (HPV)-neutralizing antibodies without antigen-specific sorting. These results validate that ImmuneAgent is a generalizable framework for rapid therapeutic antibody discovery against emerging viral threats.
ER  -