EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

GAHIB: graph attention VAE with a hyperbolic information bottleneck for biologically structured single-cell representations

Zeyu Fu, Jiawei Fu, Xiaoxia Wang, Yiyao Liu, Tianfei Ran

Frontiers in Genetics · 26.06.2026 · 17

DOI: 10.3389/fgene.2026.1863100 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Genetics, online erschienen am 26.06.2026. Verfasst von Zeyu Fu und 4 weiteren Autor:innen. Beteiligte Einrichtung: State Key Laboratory of Trauma and Chemical Poisoning, Institute of Combined Injury, Chongqing Engineering Research Center for Nanomedicine, College …. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Army Medical University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Current single-cell RNA-sequencing (scRNA-seq) variational autoencoders (VAEs) usually emphasise local cell graph structure, hyperbolic latent geometry, or bottleneck compression separately, yet these biases are rarely evaluated together in one evidence-gated representation model. We present GAHIB (graph attention VAE with a hyperbolic information bottleneck), which combines a graph-attention encoder, a 2D information bottleneck, and a Lorentz-hyperbolic geometry loss. Because the main clustering benchmark uses Leiden-derived proxy labels rather than definitive biological ground truth, we evaluate the model in two tiers: a broad 53-dataset proxy-label benchmark for method characterisation, and curated-label and marker analyses on annotated systems for biological interpretation. Across the proxy benchmark, GAHIB shows a balanced aggregate profile across clustering, projection-quality, and latent-structure metrics, while important comparisons remain mixed: scVI is statistically close to NMI/ARI, and scDHMap remains competitive on DRE-UMAP. On curated-label systems, the biological signal remains context-dependent; muscle atlas analyses support lineage-aligned structure with marker enrichment, whereas the fine T-cell immune-subtype task favors scVI. Sensitivity, seed-stability, a bounded count-dropout pilot, and cost analyses indicate practical runtime under the tested settings; however, the dropout evidence is limited to named pilot systems, and complete manually curated provenance remains an explicit limitation. Together, the results position GAHIB as a complementary, geometrically aware, single-cell representation rather than a drop-in clustering replacement.

Schlagwörter

Stichwörter der Autor:innen

graph attention networkhyperbolic geometryinformation bottleneckrepresentation learningsingle-cell RNA-seqvariational autoencoder

Autor:innen und Institutionen

  • State Key Laboratory of Trauma and Chemical Poisoning, Institute of Combined Injury, Chongqing Engineering Research Center for Nanomedicine, College of Preventive Medicine, Army Medical University, Chongqing, China.
  • Department of Orthopedics, Xinqiao Hospital, Army Medical University, Chongqing, China.
  • Department of Hepatobiliary Surgery, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China.
  • State Key Laboratory of Trauma and Chemical Poisoning, Key Laboratory of Medical Protection for Electromagnetic Radiation, Ministry of Education of China, Institute of Toxicology, College of Preventive Medicine, Army Me…

Förderung

  • Army Medical University

Zitieren

APA

Fu, Z., Fu, J., Wang, X., Liu, Y., & Ran, T. (2026). GAHIB: graph attention VAE with a hyperbolic information bottleneck for biologically structured single-cell representations. Frontiers in Genetics, 17. https://doi.org/10.3389/fgene.2026.1863100

Vancouver

Fu Z, Fu J, Wang X, Liu Y, Ran T. GAHIB: graph attention VAE with a hyperbolic information bottleneck for biologically structured single-cell representations. Frontiers in Genetics. 2026;17. doi:10.3389/fgene.2026.1863100

BibTeX
@article{Fu2026,
  author = {Fu, Zeyu and Fu, Jiawei and Wang, Xiaoxia and Liu, Yiyao and Ran, Tianfei},
  title = {GAHIB: graph attention VAE with a hyperbolic information bottleneck for biologically structured single-cell representations},
  journal = {Frontiers in Genetics},
  year = {2026},
  volume = {17},
  doi = {10.3389/fgene.2026.1863100}
}
Kostenlos starten →