Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning

• URL: http://arxiv.org/abs/2506.02911v1
• Date: Tue, 03 Jun 2025 14:16:53 GMT
• Title: Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning
• Authors: Yin Fang, Qiao Jin, Guangzhi Xiong, Bowen Jin, Xianrui Zhong, Siru Ouyang, Aidong Zhang, Jiawei Han, Zhiyong Lu,
• Abstract summary: We introduce the CellPuzzles task, where the objective is to assign unique cell types to a batch of cells.<n>This benchmark spans diverse tissues, diseases, and donor conditions, and requires reasoning across the batch-level cellular context to ensure label uniqueness.<n>We propose Cell-o1, a 7B LLM trained via supervised fine-tuning on distilled reasoning traces, followed by reinforcement learning with batch-level rewards.
• Score: 44.91329557101423
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• CellVerse: Do Large Language Models Really Understand Cell Biology? [74.34984441715517]
  We introduce CellVerse, a unified language-centric question-answering benchmark that integrates four types of single-cell multi-omics data.<n>We systematically evaluate the performance across 14 open-source and closed-source LLMs ranging from 160M to 671B on CellVerse.
  arXiv  Detail & Related papers  (2025-05-09T06:47:23Z)
• Single-Cell Omics Arena: A Benchmark Study for Large Language Models on Cell Type Annotation Using Single-Cell Data [13.56585855722118]
  Large language models (LLMs) have demonstrated their ability to efficiently process and synthesize vast corpora of text to automatically extract biological knowledge.<n>Our study explores the potential of LLMs to accurately classify and annotate cell types in single-cell RNA sequencing (scRNA-seq) data.<n>The results demonstrate that LLMs can provide robust interpretations of single-cell data without requiring additional fine-tuning.
  arXiv  Detail & Related papers  (2024-12-03T23:58:35Z)
• Cell as Point: One-Stage Framework for Efficient Cell Tracking [54.19259129722988]
  We propose a novel end-to-end one-stage framework that reimagines cell tracking by treating Cell as Point.<n>Unlike traditional methods, CAP eliminates the need for explicit detection or segmentation, instead jointly tracking cells for sequences in one stage.<n> CAP demonstrates promising cell tracking performance and is 10 to 55 times more efficient than existing methods.
  arXiv  Detail & Related papers  (2024-11-22T10:16:35Z)
• Lower-dimensional projections of cellular expression improves cell type classification from single-cell RNA sequencing [12.66369956714212]
  Single-cell RNA sequencing (scRNA-seq) enables the study of cellular diversity at single cell level. Various statistical, machine and deep learning-based methods have been proposed for cell-type classification. In this work, we proposed a reference-based method for cell type classification, called EnProCell.
  arXiv  Detail & Related papers  (2024-10-13T19:01:38Z)
• LangCell: Language-Cell Pre-training for Cell Identity Understanding [3.6518971609937068]
  We introduce LangCell, a unified representation of single-cell data and natural language during the pre-training phase. Results show that LangCell is the only single-cell PLM that can work effectively in zero-shot cell identity understanding scenarios.
  arXiv  Detail & Related papers  (2024-05-09T10:04:05Z)
• UniCell: Universal Cell Nucleus Classification via Prompt Learning [76.11864242047074]
  We propose a universal cell nucleus classification framework (UniCell) It employs a novel prompt learning mechanism to uniformly predict the corresponding categories of pathological images from different dataset domains. In particular, our framework adopts an end-to-end architecture for nuclei detection and classification, and utilizes flexible prediction heads for adapting various datasets.
  arXiv  Detail & Related papers  (2024-02-20T11:50:27Z)
• Single-cell Multi-view Clustering via Community Detection with Unknown Number of Clusters [64.31109141089598]
  We introduce scUNC, an innovative multi-view clustering approach tailored for single-cell data. scUNC seamlessly integrates information from different views without the need for a predefined number of clusters. We conducted a comprehensive evaluation of scUNC using three distinct single-cell datasets.
  arXiv  Detail & Related papers  (2023-11-28T08:34:58Z)
• Mixed Models with Multiple Instance Learning [51.440557223100164]
  We introduce MixMIL, a framework integrating Generalized Linear Mixed Models (GLMM) and Multiple Instance Learning (MIL) Our empirical results reveal that MixMIL outperforms existing MIL models in single-cell datasets.
  arXiv  Detail & Related papers  (2023-11-04T16:42:42Z)
• OCELOT: Overlapped Cell on Tissue Dataset for Histopathology [13.691924123273004]
  We release OCELOT, a dataset dedicated to the study of cell-tissue relationships for cell detection in histopathology. We propose multi-task learning approaches that benefit from learning both cell and tissue tasks simultaneously. On the OCELOT test set in particular, we show up to 6.79 improvement in F1-score.
  arXiv  Detail & Related papers  (2023-03-23T08:57:11Z)
• Split and Expand: An inference-time improvement for Weakly Supervised Cell Instance Segmentation [71.50526869670716]
  We propose a two-step post-processing procedure, Split and Expand, to improve the conversion of segmentation maps to instances. In the Split step, we split clumps of cells from the segmentation map into individual cell instances with the guidance of cell-center predictions. In the Expand step, we find missing small cells using the cell-center predictions.
  arXiv  Detail & Related papers  (2020-07-21T14:05:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.