CyteTypeR

Automated, evidence-based cell type annotation for single-cell transcriptomics

🎉 NEW: Preprint published November 7, 2025 on bioRxiv
📅 FREE Webinar: Register now — Learn CyteType from the developers

Why CyteTypeR?

Manual cell type annotation takes weeks and varies between experts. CyteTypeR delivers consistent, expert-level annotations in minutes using a multi-agent AI system where specialized agents collaborate on marker analysis, literature evidence, and Cell Ontology mapping.

• Save weeks of manual curation — Annotate entire datasets at expert level in minutes, not days
• Drop-in integration — 3 lines of code, works with existing Scanpy/Seurat workflows
• No setup friction — No API keys required; built-in LLM with optional custom configuration
• Standards-compliant output — Automatic Cell Ontology term mapping (CL IDs)
• Comprehensive annotations — Cell types, subtypes, activation states, confidence scores, and lineage
• Transparent and auditable — Interactive HTML reports show evidence, reasoning, and confidence for every annotation

See example report

Installation

# Using devtools
install.packages("devtools")

# Install from GitHub
library(devtools)
install_github("NygenAnalytics/CyteTypeR")

Quick Start

# Load package
library(CyteTypeR)

prepped_data <- PrepareCyteTypeR(
  pbmc,
  pbmc.markers,
  n_top_genes = 10,
  group_key = 'seurat_clusters',
  aggregate_metadata = TRUE,
  coordinates_key = "umap"
)

metadata <- list(
  title = 'My scRNA-seq analysis of human pbmc',
  run_label = 'initial_analysis',
  experiment_name = 'pbmc_human_samples_study'
)

results <- CyteTypeR(
  obj=pbmc,
  prepped_data = prepped_data, 
  study_context = "pbmc blood samples from humans", 
  metadata = metadata
)

Note: No API keys required for default configuration. See custom LLM configuration for advanced options.

Using Scanpy/Anndata? → CyteType

Output Reports

Each analysis generates an HTML report documenting annotation decisions, marker genes, confidence scores, and Cell Ontology mappings:

View example report with embedded chat interface

Benchmarks

Validated across multiple datasets, tissues, and organisms. CyteType's agentic architecture consistently outperforms other methods across multiple LLMs:

📊 Performance: 388% improvement over GPTCellType, 268% over CellTypist, 101% over SingleR

Browse results from single-cell atlases →

Need Help?

📖 Configuration options 💬 Join Discord for support

Citation

If you use CyteTypeR in your research, please cite our preprint:

@article{cytetype2025,
  title={Multi-agent AI enables evidence-based cell annotation in single-cell transcriptomics},
  author={Gautam Ahuja, Alex Antill, Yi Su, Giovanni Marco Dall'Olio, Sukhitha Basnayake, Göran Karlsson, Parashar Dhapola},
  journal={bioRxiv},
  year={2025},
  doi={10.1101/2025.11.06.686964},
  url={https://www.biorxiv.org/content/10.1101/2025.11.06.686964v1}
}

License

CyteTypeR is free for academic and non-commercial research use under CC BY‑NC‑SA 4.0 — see LICENSE.md

For commercial use, please contact us at [email protected]