Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912411520335872 |
|---|---|
| author | Fang, Yin Jin, Qiao Xiong, Guangzhi Jin, Bowen Zhong, Xianrui Ouyang, Siru Zhang, Aidong Han, Jiawei Lu, Zhiyong |
| author_facet | Fang, Yin Jin, Qiao Xiong, Guangzhi Jin, Bowen Zhong, Xianrui Ouyang, Siru Zhang, Aidong Han, Jiawei Lu, Zhiyong |
| contents | Cell type annotation is a key task in analyzing the heterogeneity of single-cell RNA sequencing data. Although recent foundation models automate this process, they typically annotate cells independently, without considering batch-level cellular context or providing explanatory reasoning. In contrast, human experts often annotate distinct cell types for different cell clusters based on their domain knowledge. To mimic this workflow, we introduce the CellPuzzles task, where the objective is to assign unique cell types to a batch of cells. This benchmark spans diverse tissues, diseases, and donor conditions, and requires reasoning across the batch-level cellular context to ensure label uniqueness. We find that off-the-shelf large language models (LLMs) struggle on CellPuzzles, with the best baseline (OpenAI's o1) achieving only 19.0% batch-level accuracy. To fill this gap, we propose Cell-o1, a 7B LLM trained via supervised fine-tuning on distilled reasoning traces, followed by reinforcement learning with batch-level rewards. Cell-o1 achieves state-of-the-art performance, outperforming o1 by over 73% and generalizing well across contexts. Further analysis of training dynamics and reasoning behaviors provides insights into batch-level annotation performance and emergent expert-like reasoning. Code and data are available at https://github.com/ncbi-nlp/cell-o1. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_02911 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning Fang, Yin Jin, Qiao Xiong, Guangzhi Jin, Bowen Zhong, Xianrui Ouyang, Siru Zhang, Aidong Han, Jiawei Lu, Zhiyong Computation and Language Artificial Intelligence Computational Engineering, Finance, and Science Human-Computer Interaction Machine Learning Cell type annotation is a key task in analyzing the heterogeneity of single-cell RNA sequencing data. Although recent foundation models automate this process, they typically annotate cells independently, without considering batch-level cellular context or providing explanatory reasoning. In contrast, human experts often annotate distinct cell types for different cell clusters based on their domain knowledge. To mimic this workflow, we introduce the CellPuzzles task, where the objective is to assign unique cell types to a batch of cells. This benchmark spans diverse tissues, diseases, and donor conditions, and requires reasoning across the batch-level cellular context to ensure label uniqueness. We find that off-the-shelf large language models (LLMs) struggle on CellPuzzles, with the best baseline (OpenAI's o1) achieving only 19.0% batch-level accuracy. To fill this gap, we propose Cell-o1, a 7B LLM trained via supervised fine-tuning on distilled reasoning traces, followed by reinforcement learning with batch-level rewards. Cell-o1 achieves state-of-the-art performance, outperforming o1 by over 73% and generalizing well across contexts. Further analysis of training dynamics and reasoning behaviors provides insights into batch-level annotation performance and emergent expert-like reasoning. Code and data are available at https://github.com/ncbi-nlp/cell-o1. |
| title | Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning |
| topic | Computation and Language Artificial Intelligence Computational Engineering, Finance, and Science Human-Computer Interaction Machine Learning |
| url | https://arxiv.org/abs/2506.02911 |