CellFM: a large-scale foundation model pre-trained on transcriptomics of 100 million human cells

Single-cell sequencing provides transcriptomic profiling at single-cell resolution, uncovering cellular heterogeneity with unprecedented precision. Yet, current single cell data analysis suffers from the inherent data noises, batch effects, and sparsity, highlighting the requirement of a unified model to represent cellular states. To circumvent this problem, many recent efforts focus on training single-cell foundation models based on large datasets. However, current human foundation models are still limited by the sizes of training data and model parameters. Here, we have collected a diverse dataset of 100 million human cells, on which we train a single-cell foundation model (CellFM) containing 800 million parameters. To balance efficiency and performance, the model is trained through a modified RetNet framework on the MindSpore. Extensive experiments have shown that CellFM outperforms existing models in cell annotation, perturbation prediction, gene function prediction, and gene-gene relationship capturing.

Authors

Institutions

Publication Details

Journal
Nature Communications
Published
2025-05-20
DOI
https://doi.org/10.1038/s41467-025-59926-5
Citations
35
Primary Topic
Single-cell and spatial transcriptomics
Type
article
Field-Weighted Citation Impact
28.21

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

CellFM: a large-scale foundation model pre-trained on transcriptomics of 100 million human cells

Wenbing Li, Zhuoyi Wei, Yuedong Yang, Huiying Zhao et al.
35 citations
Nature Communications
Single-cell and spatial transcriptomics
28.21
article

CellFM: a large-scale foundation model pre-trained on transcriptomics of 100 million human cells

Wenbing Li, Zhuoyi Wei, Yuedong Yang, Huiying Zhao, Shuangyu Yang, Weijiang Yu, Chengyang Zhang, Jinbo Zhang, Ningyuan Shangguan, Hongyu Zhang, Jiancong Xie, Yutong Lu, Nan Fang, Yuansong Zeng, Yun Su, Jue Fan
article en
35 citations

Abstract

Single-cell sequencing provides transcriptomic profiling at single-cell resolution, uncovering cellular heterogeneity with unprecedented precision. Yet, current single cell data analysis suffers from the inherent data noises, batch effects, and sparsity, highlighting the requirement of a unified model to represent cellular states. To circumvent this problem, many recent efforts focus on training single-cell foundation models based on large datasets. However, current human foundation models are still limited by the sizes of training data and model parameters. Here, we have collected a diverse dataset of 100 million human cells, on which we train a single-cell foundation model (CellFM) containing 800 million parameters. To balance efficiency and performance, the model is trained through a modified RetNet framework on the MindSpore. Extensive experiments have shown that CellFM outperforms existing models in cell annotation, perturbation prediction, gene function prediction, and gene-gene relationship capturing.

Nature CommunicationsVol. 16(1)
Sun Yat-sen University (CN), Chongqing University (CN), Huawei Technologies (China) (CN), Sun Yat-sen Memorial Hospital (CN), Singleron Biotechnologies (china) (CN)
National Science Foundation, National Natural Science Foundation of China, China Postdoctoral Science Foundation, Government of Jiangsu Province, Fundamental Research Funds for the Central Universities
Openalex Percentile: Top 0%
Single-cell and spatial transcriptomics
28.21
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.