EUVIMEDEUVIMED entdecken
Open Access

EpiMII: Structure-Aware Graph Neural Networks for MHC-II Epitope Generation

Jiayi Yuan, Xiaowei Xu, Ze-Yu Sun, Shan Zhu, Chunyu Wei, Tianjian Liang, Jingxuan Ge, Xiang-Qun Xie, Yan Chen, Zhiwei Feng, Tingjun Hou iD, Ying Xue

Research · 01.01.2026 · 9

DOI: 10.34133/research.1311 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Research, online erschienen am 15.06.2026. Verfasst von Jiayi Yuan und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Pharmaceutical Sciences, Computational Chemical Genomics Screening Center, and Pharmacometrics & System Pharmacology PharmacoAnalytics,…. Die Arbeit nennt 67 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIH NIDA.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Major histocompatibility complex class II (MHC-II) neoantigens play a critical role in immunotherapy, either as direct effectors or through their influence on CD8 + T cells. However, only a small fraction of tumor DNA mutations qualify as functional neoantigens, and current prediction tools suffer from limited accuracy, making the in vivo design of highly immunogenic neoantigens challenging. Here, we present EpiMII, a structure-aware graph neural network model for de novo design of MHC-II epitopes. By integrating 3-dimensional structural information through an inverse folding strategy, EpiMII generates mimotopes that preserve T cell specificity while enhancing MHC-II binding affinity and immunogenicity. The model was trained on a first-of-its-kind curated dataset of 142,934 homology-modeled MHC-II epitope structures, overcoming the scarcity of experimentally resolved MHC-II–peptide complexes. Benchmarking analyses illustrate that EpiMII achieves a sequence recovery rate of 66.7% on a held-out test set and 79.0% on crystallized epitopes, markedly outperforming existing tools like ProteinMPNN. In a hepatocellular carcinoma case study, all 5 EpiMII-designed epitopes could markedly activate CD4 + T cells in vitro and induce the secretion of interferon-γ and tumor necrosis factor-α. Notably, treatment with one epitope (P4) substantially reduced tumor volume in mice in vivo, comparable to a known positive control. These findings establish EpiMII as a powerful tool for structure-guided neoantigen discovery, with broad implications for cancer vaccine development and personalized immunotherapy.

Autor:innen und Institutionen

  • Department of Pharmaceutical Sciences, Computational Chemical Genomics Screening Center, and Pharmacometrics & System Pharmacology PharmacoAnalytics, School of Pharmacy; National Center of Excellence for Computational D…
  • College of Pharmacology Sciences, Zhejiang University of Technology, Hangzhou 313099, Zhejiang, China.
  • Faculty of Pharmaceutical Sciences, Shenzhen University of Advanced Technology, Shenzhen 518107, Guangdong, China.
  • College of Pharmaceutical Sciences, Zhejiang University, Hangzhou 310058, Zhejiang, China.
  • Department of Pharmacy; Center of Excellence for Intelligent Pharmacy and Precision Pharmacotherapy (CE-IPPP); Shanghai International Medical Innovation Center (SIMIC); National Medical Center, Zhongshan Hospital, Fudan…

ORCID-Profile: Tingjun Hou

Förderung

  • NIH NIDA

Zitieren

APA

Yuan, J., Xu, X., Sun, Z. Y., Zhu, S., Wei, C., Liang, T., Ge, J., Xie, X. Q., Chen, Y., Feng, Z., Hou, T., & Xue, Y. (2026). EpiMII: Structure-Aware Graph Neural Networks for MHC-II Epitope Generation. Research, 9. https://doi.org/10.34133/research.1311

Vancouver

Yuan J, Xu X, Sun ZY, Zhu S, Wei C, Liang T, et al. EpiMII: Structure-Aware Graph Neural Networks for MHC-II Epitope Generation. Research. 2026;9. doi:10.34133/research.1311

BibTeX
@article{Yuan2026,
  author = {Yuan, Jiayi and Xu, Xiaowei and Sun, Ze-Yu and Zhu, Shan and Wei, Chunyu and Liang, Tianjian and Ge, Jingxuan and Xie, Xiang-Qun and Chen, Yan and Feng, Zhiwei and Hou, Tingjun and Xue, Ying},
  title = {EpiMII: Structure-Aware Graph Neural Networks for MHC-II Epitope Generation},
  journal = {Research},
  year = {2026},
  volume = {9},
  doi = {10.34133/research.1311}
}
Kostenlos starten →