Skip to content

Commit 4cdf613

Browse files
parashardhapolaparashardhapola
authored
Revise README for clarity and new information
Updated README to enhance clarity on CyteType features and added new sections for quick start, benchmarks, and citation.
1 parent d9b2689 commit 4cdf613

File tree

1 file changed

+80
-49
lines changed

1 file changed

+80
-49
lines changed

‎README.md‎

Lines changed: 80 additions & 49 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
<h1 align="left">CyteType</h1>
2-
<h4 align="left">Multi-agent AI driven cell annotation in single-cell transcriptomics</h4>
2+
<h3 align="left">Automated, evidence-based cell type annotation for single-cell transcriptomics</h3>
33

44
<p align="left">
55
<a href="https://github.com/NygenAnalytics/cytetype/actions/workflows/publish.yml">
@@ -15,69 +15,100 @@
1515
<a href="https://pypi.org/project/cytetype/">
1616
<img src="https://img.shields.io/pypi/dm/cytetype" alt="PyPI downloads">
1717
</a>
18-
<a href="https://discord.gg/V6QFM4AN">
19-
<img src="https://img.shields.io/discord/1339594966300622908" alt="PyPI version">
20-
</a>
2118
</p>
2219

2320
---
24-
<a href="https://colab.research.google.com/drive/1aRLsI3mx8JR8u5BKHs48YUbLsqRsh2N7?usp=sharing" target="_blank">Example Notebook</a> |
25-
<a href="https://prod.cytetype.nygen.io/report/6420a807-8bf3-4c33-8731-7617edfc2ad0?v=251124" target="_blank">Example output</a> |
26-
<a href="docs/examples.md">Atlas scale results</a>
2721

28-
Switch to R/Seurat package: <a href="https://github.com/NygenAnalytics/CyteTypeR">CyteTypeR</a>
22+
> 🎉 **NEW:** [Preprint published November 7, 2025](https://www.biorxiv.org/content/10.1101/2025.11.06.686964v1) on bioRxiv
23+
> 📅 **FREE Webinar:** [Register now](https://attendee.gotowebinar.com/register/1731194703386732893) — Learn CyteType from the developers
24+
25+
---
26+
27+
## Why CyteType?
28+
29+
Manual cell type annotation takes weeks and varies between experts. CyteType 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.
2930

30-
Read Preprint: <a href="https://www.biorxiv.org/content/10.1101/2025.11.06.686964v1">BioRxiv</a>
31+
<img width="800" alt="CyteType Overview" src="https://github.com/user-attachments/assets/c4cc4f67-9c63-4590-9717-c2391b3e5faf" />
3132

32-
## Installation
33-
`pip install cytetype`
33+
- **Save weeks of manual curation** — Annotate entire datasets at expert level in minutes, not days
34+
- **Drop-in integration** — 3 lines of code, works with existing Scanpy/Seurat workflows
35+
- **No setup friction** — No API keys required; built-in LLM with optional custom configuration
36+
- **Standards-compliant output** — Automatic Cell Ontology term mapping (CL IDs)
37+
- **Comprehensive annotations** — Cell types, subtypes, activation states, confidence scores, and lineage
38+
- **Transparent and auditable** — Interactive HTML reports show evidence, reasoning, and confidence for every annotation
39+
40+
**🚀 [Try it now in Colab](https://colab.research.google.com/drive/1aRLsI3mx8JR8u5BKHs48YUbLsqRsh2N7?usp=sharing)** • **[See example report](https://prod.cytetype.nygen.io/report/6420a807-8bf3-4c33-8731-7617edfc2ad0?v=251124)**
41+
42+
---
3443

3544
## Quick Start
3645

46+
```bash
47+
pip install cytetype
48+
```
49+
3750
```python
38-
import anndata
3951
import scanpy as sc
4052
from cytetype import CyteType
4153

42-
# ------ Example Scanpy Pipeline ------
43-
# Skip this step if you already have clusters and marker genes in an AnnData object.
44-
adata = anndata.read_h5ad("path/to/your/data.h5ad")
45-
sc.pp.normalize_total(adata, target_sum=1e4)
46-
sc.pp.log1p(adata)
47-
sc.pp.highly_variable_genes(adata, n_top_genes=1000)
48-
sc.pp.pca(adata)
49-
sc.pp.neighbors(adata)
50-
51-
group_key = 'clusters' # Wherever you want to store or already have clusters in adata.obs
52-
53-
sc.tl.leiden(adata, key_added=group_key)
54-
sc.tl.umap(adata)
55-
sc.tl.rank_genes_groups(adata, groupby=group_key, method="t-test")
56-
# ------ Example Scanpy Pipeline ------
57-
58-
# ------ CyteType ------
59-
annotator = CyteType(adata, group_key=group_key)
60-
adata = annotator.run(
61-
study_context="Brief study description (e.g., Human brain tissue ...)",
62-
)
63-
64-
# Visualize results
65-
sc.pl.embedding(adata, basis='umap', color=f'cytetype_annotation_{group_key}')
66-
sc.pl.embedding(adata, basis='umap', color=f'cytetype_cellOntologyTerm_{group_key}')
67-
sc.pl.embedding(adata, basis='umap', color=f'cytetype_cellOntologyTermID_{group_key}')
68-
sc.pl.embedding(adata, basis='umap', color=f'cytetype_cellState_{group_key}')
54+
# Assumes preprocessed AnnData with clusters and marker genes
55+
annotator = CyteType(adata, group_key='clusters')
56+
adata = annotator.run(study_context="Human PBMC from healthy donor")
57+
sc.pl.umap(adata, color='cytetype_annotation_clusters')
6958
```
7059

71-
## Documentation
72-
- Configuration (LLMs, auth, advanced): <a href="docs/configuration.md">docs/configuration.md</a>
73-
- Results: <a href="docs/results.md">docs/results.md</a>
74-
- Troubleshooting: <a href="docs/troubleshooting.md">docs/troubleshooting.md</a>
75-
- Development: <a href="docs/development.md">docs/development.md</a>
76-
- Server Overview (high‑level): <a href="docs/server-overview.md">docs/server-overview.md</a>
77-
- Ollama Integration: <a href="docs/ollama.md">docs/ollama.md</a>
60+
> **Note:** No API keys required for default configuration. See [custom LLM configuration](docs/configuration.md#llm-configuration) for advanced options.
7861
79-
## License
80-
Licensed under CC BY‑NC‑SA 4.0 — see <a href="LICENSE.md">LICENSE.md</a>.
62+
**Using R/Seurat?** → [CyteTypeR](https://github.com/NygenAnalytics/CyteTypeR)
63+
64+
---
65+
66+
## Output Reports
67+
68+
Each analysis generates an HTML report documenting annotation decisions, marker genes, confidence scores, and Cell Ontology mappings:
69+
70+
<img width="1000" alt="CyteType Report Example" src="https://github.com/user-attachments/assets/9f0f4b36-2dd7-4cb8-93e3-ecda9c97a930" />
71+
72+
**[View example report with embedded chat interface](https://prod.cytetype.nygen.io/report/6420a807-8bf3-4c33-8731-7617edfc2ad0?v=251124)**
73+
74+
---
75+
76+
## Benchmarks
77+
78+
Validated across multiple datasets, tissues, and organisms. CyteType's agentic architecture consistently outperforms other methods across multiple LLMs:
79+
80+
**📊 Performance:** 388% improvement over GPTCellType, 268% over CellTypist, 101% over SingleR
81+
82+
<img width="500" alt="CyteType Benchmark Results" src="https://github.com/user-attachments/assets/a63cadc1-d8c5-4ac0-bba7-af36f9b3c46d" />
83+
84+
**[Browse results from single-cell atlases →](docs/examples.md)**
85+
86+
## Need Help?
87+
88+
📖 [Configuration options](docs/configuration.md) • [Understanding output columns](docs/results.md) • [Common issues & fixes](docs/troubleshooting.md) • [Development guide](docs/development.md)
89+
💬 [Join Discord](https://discord.gg/V6QFM4AN) for support
90+
91+
---
92+
93+
## Citation
94+
95+
If you use CyteType in your research, please cite our preprint:
96+
97+
```bibtex
98+
@article{cytetype2025,
99+
title={Multi-agent AI enables evidence-based cell annotation in single-cell transcriptomics},
100+
author={Gautam Ahuja, Alex Antill, Yi Su, Giovanni Marco Dall'Olio, Sukhitha Basnayake, Göran Karlsson, Parashar Dhapola},
101+
journal={bioRxiv},
102+
year={2025},
103+
doi={10.1101/2025.11.06.686964},
104+
url={https://www.biorxiv.org/content/10.1101/2025.11.06.686964v1}
105+
}
106+
```
81107

82108
---
83-
<img width="2063" height="1857" alt="CyteType architecture" src="https://github.com/user-attachments/assets/c55f00a2-c4d1-420a-88c2-cdb507898383" />
109+
110+
## License
111+
112+
CyteType is free for academic and non-commercial research use under CC BY‑NC‑SA 4.0 — see [LICENSE.md](LICENSE.md)
113+
114+
For commercial use, please contact us at [contact@nygen.io](mailto:contact@nygen.io)

0 commit comments

Comments
 (0)