EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A graph-attentive GAN for rare-cell-aware single-cell RNA-seq data generation

Ritwik Ganguly iD, Sana Aafrine iD, Sk Md Mosaddek Hossain iD, Sumanta Ray iD

PLOS Computational Biology · 2026

DOI: 10.1371/journal.pcbi.1014601 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Computational Biology, 2026. Verfasst von Ritwik Ganguly und 3 weiteren Autor:innen. Beteiligte Einrichtung: Computer Science and Engineering, Aliah University, Kolkata, India. Hauptthemen laut MEDLINE: Single-Cell Analysis, RNA-Seq, Sequence Analysis, RNA. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

A central challenge in downstream single-cell RNA sequencing (scRNA-seq) analysis is the high-dimensional, small-sample (HDSS) regime, often compounded by class imbalance from rare cell types. These factors hinder robust feature (gene) selection and cell clustering and limit the realism of samples generated by existing simulators. We introduce GARAGE , a G raph- A ttentive RA re-cell aware single-cell data GE neration that augments the generator’s input with a small, attention-weighted ‘leakage’ of real cells in addition to prior noise. Specifically, we build a k -nearest-neighbour cell graph and use a graph attention network (GAT) to prioritize nodes that likely represent under-sampled (rare) subpopulations; these high-attention cell embeddings are injected into the generator input to steer synthesis toward biologically plausible regions of the data manifold while respecting cell-type proportions. This attention-guided leakage accelerates training, reduces mode dropping, and yields realistic synthetic cells that preserve rare-cell structure. Across real scRNA-seq benchmarks, GARAGE improves downstream feature selection and clustering compared with state-of-the-art baselines. In summary, GARAGE directly addresses HDSS and rarity in scRNA-seq by coupling graph attention with adversarial generation to produce high-fidelity synthetic cells that enhance downstream analyses. The corresponding software is available at: https://github.com/RitwikGanguly/GARAGE .

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Single-Cell AnalysisRNA-SeqSequence Analysis, RNASingle-Cell Gene Expression AnalysisComputational BiologyGenerative Adversarial NetworksAlgorithmsGraph Neural NetworksAnimalsHumansSoftware

Autor:innen und Institutionen

  • Computer Science and Engineering, Aliah University, Kolkata, India.
  • Computational Biology, Indraprastha Institute of Information Technology, New Delhi, India.
  • Data Science, The West Bengal National University of Juridical Sciences, Kolkata, India.

ORCID-Profile: Ritwik Ganguly, Sana Aafrine, Sk Md Mosaddek Hossain, Sumanta Ray

Zitieren

APA

Ganguly, R., Aafrine, S., Hossain, S. M. M., & Ray, S. (2026). A graph-attentive GAN for rare-cell-aware single-cell RNA-seq data generation. PLOS Computational Biology. https://doi.org/10.1371/journal.pcbi.1014601

Vancouver

Ganguly R, Aafrine S, Hossain SMM, Ray S. A graph-attentive GAN for rare-cell-aware single-cell RNA-seq data generation. PLOS Computational Biology. 2026. doi:10.1371/journal.pcbi.1014601

BibTeX
@article{Ganguly2026,
  author = {Ganguly, Ritwik and Aafrine, Sana and Hossain, Sk Md Mosaddek and Ray, Sumanta},
  title = {A graph-attentive GAN for rare-cell-aware single-cell RNA-seq data generation},
  journal = {PLOS Computational Biology},
  year = {2026},
  doi = {10.1371/journal.pcbi.1014601}
}
Kostenlos starten →