---
license: mit
language:
  - en
tags:
  - biology
  - single-cell
  - scFM
  - Drosophila
  - foundation-model
  - cell-type-annotation
---

<div align="start">
  <img src="lemur.png" alt="Lemur Icon" width="250">
  <h1>Lemur: A Single-Cell Foundation Model for <em>Drosophila melanogaster</em> 🪰🔬</h1>
</div>

**Lemur** is a single-cell foundation model developed specifically for *Drosophila melanogaster*. It achieves fine-tuning-free cell-type annotation using two key innovations:

- **Comprehensive Pre-training:** Trained on an integrated whole-organism atlas with a unified cell-type annotation schema.
- **Hierarchical Decoding:** Employs a transformer-based hierarchical cell-type decoder to generate consistent predictions at multiple levels of granularity.

### Highlights ⚡️

- **Fine-Tuning-Free Annotation:** No extra training required on new datasets.
- **Robust Across Conditions:** Consistent performance across various tissue types, experimental conditions, and sequencing platforms.
- **Zero-Shot Batch-Effect Correction:** Automatically corrects batch effects without further adjustments.
- **Broad Applications:** Tailored for the fly research community with potential implications for aging and neurodegenerative disease studies in humans.

### Status & Availability 🚀

- **Model Weights:** Available now on the Hugging Face Hub 🏋🏻‍♂️.
- **Code:** Soon to be released. Stay tuned! 🔥👀

### Publication & Citation 📚🔗

We're excited to share our manuscript, *Lemur: A Single-Cell Foundation Model with Fine-Tuning-Free Hierarchical Cell-Type Generation for* *Drosophila melanogaster*, which is now available as a preprint on [bioRxiv: 10.1101/123456](https://www.biorxiv.org/content/10.1101/123456)! 🚀✨

If you find our work helpful, please consider citing our preprint as follows:

```bibtex
@article{your_citation_key,
  title={Lemur: A Single-Cell Foundation Model with Fine-Tuning-Free Hierarchical Cell-Type Generation for *Drosophila melanogaster*},
  author={Doe, John and Smith, Jane and Collaborators},
  journal={bioRxiv},
  year={2025},
  doi={10.1101/123456}
}
```