Jue Fan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Jue Fan has a total of 70 co-authors. Here are the top co-authors by number of shared publications:
- Xiao Sun
Affiliation State Key Laboratory of Digital Medical Engineering, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China
3 shared publications Yan Zhang
Singleron Biotechnologies
3 shared publicationsJinbo Zhang
Singleron Biotechnologies, Nanjing, Jiangsu, China
2 shared publicationsLongfei Hu
Singleron Biotechnologies, Yaogu Avenue 11, Nanjing, Jiangsu, China
2 shared publications- Nan Fang
Singleron Biotechnologies
2 shared publications - Nan Fang
Singleron Biotechnologies, Nanjing, Jiangsu, China
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 20 May 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-17
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.
- 20 May 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-17
Top Cited PapersOpen AccessCellFM: 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.
- Open Access
- 12 May 2025
- Published by EDP Sciences in BIO Web of Conferences
- Vol. 174, 03016
ScBlkCom: An Integrated Compression Algorithm for Single-Cell RNA Sequencing Data
High-throughput sequencing advancements have shifted genomic project bottlenecks from data generation to computational storage and analysis. Single-cell RNA-seq (scRNA-seq) data exhibits unique structural features, including extensive labeled sequence identifiers, which conventional compression tools fail to optimize. This study proposes ScBlkCom, a specialized compression scheme for scRNA-seq data. The method partitions sequencing data into distinct blocks and applies tailored compression strategies: differential encoding for numerical attributes, Huffman coding for categorical labels, and context-adaptive encoding for sequence identifiers. Experiments demonstrate ScBlkCom achieves 84.29% higher compression gain compared to single-module approaches and outperforms generic tools (e.g., GZIP, BZIP2) by 6.44% in compression ratio, while maintaining stable processing speeds. This block-wise adaptive framework effectively addresses scRNA-seq data redundancy, offering enhanced storage efficiency for large-scale single-cell studies.
- 12 May 2025
- Published by EDP Sciences in BIO Web of Conferences
- Vol. 174, 03016
Open AccessScBlkCom: An Integrated Compression Algorithm for Single-Cell RNA Sequencing Data
High-throughput sequencing advancements have shifted genomic project bottlenecks from data generation to computational storage and analysis. Single-cell RNA-seq (scRNA-seq) data exhibits unique structural features, including extensive labeled sequence identifiers, which conventional compression tools fail to optimize. This study proposes ScBlkCom, a specialized compression scheme for scRNA-seq data. The method partitions sequencing data into distinct blocks and applies tailored compression strategies: differential encoding for numerical attributes, Huffman coding for categorical labels, and context-adaptive encoding for sequence identifiers. Experiments demonstrate ScBlkCom achieves 84.29% higher compression gain compared to single-module approaches and outperforms generic tools (e.g., GZIP, BZIP2) by 6.44% in compression ratio, while maintaining stable processing speeds. This block-wise adaptive framework effectively addresses scRNA-seq data redundancy, offering enhanced storage efficiency for large-scale single-cell studies.
- Open Access
- 12 May 2025
- Published by EDP Sciences in BIO Web of Conferences
- Vol. 174, 01015
Identification of Biomarkers for Inflammatory Bowel Disease Based on Single-Cell RNA Sequencing Data
Inflammatory bowel disease (IBD) is a chronic gastrointestinal disease that imposes a severe health burden globally. Single-cell RNA sequencing (scRNA-seq) technology helps to elucidate the molecular characteristics and functional states of individual cells, thereby advancing the study of IBD's intricate mechanisms. Most current studies focus on small-scale samples from a few cell subclusters, lacking large-scale, systematic analyses of IBD. By collecting and integrating multiple datasets, large-scale sample datasets were analyzed for cell type identification and reclassification, addressing this gap. In this study, we collected and merged several public IBD-related scRNA-seq datasets, creating a dataset with millions of cells. Using data mining tools and bioinformatics techniques, we have identified Treg cells, Th17 cells, the macrophage C12 subcluster, and the fibroblast C5 subcluster as being significantly associated with IBD. These cell types play crucial roles in inflammation, fibrosis, and immune regulation. Differential gene expression analysis revealed several potential biomarkers, including IL2RA, HSF1, and TNFSF8 in Treg cells; PIM3, RIPK2, and TLR2 in macrophage C12 subcluster; and CDH11, OSMR, and BRD4 in fibroblast C5 subcluster. These biomarkers could serve as potential therapeutic targets, contributing to a deeper understanding and more effective treatment of IBD.
- 12 May 2025
- Published by EDP Sciences in BIO Web of Conferences
- Vol. 174, 01015
Open AccessIdentification of Biomarkers for Inflammatory Bowel Disease Based on Single-Cell RNA Sequencing Data
Inflammatory bowel disease (IBD) is a chronic gastrointestinal disease that imposes a severe health burden globally. Single-cell RNA sequencing (scRNA-seq) technology helps to elucidate the molecular characteristics and functional states of individual cells, thereby advancing the study of IBD's intricate mechanisms. Most current studies focus on small-scale samples from a few cell subclusters, lacking large-scale, systematic analyses of IBD. By collecting and integrating multiple datasets, large-scale sample datasets were analyzed for cell type identification and reclassification, addressing this gap. In this study, we collected and merged several public IBD-related scRNA-seq datasets, creating a dataset with millions of cells. Using data mining tools and bioinformatics techniques, we have identified Treg cells, Th17 cells, the macrophage C12 subcluster, and the fibroblast C5 subcluster as being significantly associated with IBD. These cell types play crucial roles in inflammation, fibrosis, and immune regulation. Differential gene expression analysis revealed several potential biomarkers, including IL2RA, HSF1, and TNFSF8 in Treg cells; PIM3, RIPK2, and TLR2 in macrophage C12 subcluster; and CDH11, OSMR, and BRD4 in fibroblast C5 subcluster. These biomarkers could serve as potential therapeutic targets, contributing to a deeper understanding and more effective treatment of IBD.
- Open Access
CellClear: Enhancing Single-cell RNA Data Quality via Biologically-Informed Ambient RNA Correction
Ambient RNA, caused by the pool of mRNA molecules released from lysed or dead cells in samples, poses persistent challenges to single-cell RNA sequencing data analysis, and leads to potential problems such as inaccurate classification of cell types and downstream functional analyses. Here we propose CellClear to estimate and remove the ambient RNA expression programs in single-cell RNA sequencing data. CellClear demonstrates significant increase in the accuracy of ambient gene expression correction comparing with state-of-art methods, without distorting native information of individual cell types.
Open AccessCellClear: Enhancing Single-cell RNA Data Quality via Biologically-Informed Ambient RNA Correction
Ambient RNA, caused by the pool of mRNA molecules released from lysed or dead cells in samples, poses persistent challenges to single-cell RNA sequencing data analysis, and leads to potential problems such as inaccurate classification of cell types and downstream functional analyses. Here we propose CellClear to estimate and remove the ambient RNA expression programs in single-cell RNA sequencing data. CellClear demonstrates significant increase in the accuracy of ambient gene expression correction comparing with state-of-art methods, without distorting native information of individual cell types.
- Open Access
SynEcoSys: a multifunctional platform of large-scale single-cell omics data analysis
Next generation sequencing technologies enable the analysis of the transcriptomes of individual cells, providing a higher resolution of gene expression and function at the single cell level. Various single-cell data are continuously generated every year, covering fields from scientific research to clinical development. The fast-growing public datasets are collected by distinctive platforms, which are designed to facilitate biological discoveries, disease diagnosis, and new treatments. However, these platforms are hard to meet the urgency of having a unified data integration pipeline to improve comparability between datasets. Here we present SynEcoSys, an online multifunctional platform for single-cell transcriptomic data analysis, visualization, and exploration. SynEcoSys by Singleron Biotechnologies currently provides a massive collection of publicly available single-cell sequencing dataset, involving 46,326,175 cells from 731 datasets across multiple platforms and species. All datasets are generated with a strict and uniform data analysis pipeline and cell marker-based manual annotation, thus facilitating more comprehensive and reliable data mining. SynEcoSys is available at: https://www.synecosys.com .
Open AccessSynEcoSys: a multifunctional platform of large-scale single-cell omics data analysis
Next generation sequencing technologies enable the analysis of the transcriptomes of individual cells, providing a higher resolution of gene expression and function at the single cell level. Various single-cell data are continuously generated every year, covering fields from scientific research to clinical development. The fast-growing public datasets are collected by distinctive platforms, which are designed to facilitate biological discoveries, disease diagnosis, and new treatments. However, these platforms are hard to meet the urgency of having a unified data integration pipeline to improve comparability between datasets. Here we present SynEcoSys, an online multifunctional platform for single-cell transcriptomic data analysis, visualization, and exploration. SynEcoSys by Singleron Biotechnologies currently provides a massive collection of publicly available single-cell sequencing dataset, involving 46,326,175 cells from 731 datasets across multiple platforms and species. All datasets are generated with a strict and uniform data analysis pipeline and cell marker-based manual annotation, thus facilitating more comprehensive and reliable data mining. SynEcoSys is available at: https://www.synecosys.com .
- 20 September 2020
- Published by Springer Nature
- Vol. 1255, 109-121
Single Cell Sequencing: A New Dimension in Cancer Diagnosis and Treatment
Cancer is one of the leading causes of death worldwide and well known for its complexity. Cancer cells within the same tumor or from different tumors are highly heterogeneous. Furthermore, stromal and immune cells within tumor microenvironment interact with cancer cells to play important roles in how tumors progress and respond to different treatments. Recent advances in single cell technologies, especially massively parallel single cell sequencing, have made it possible to analyze cancer cells and cells in its tumor microenvironment in parallel with unprecedented high resolution. In this chapter, we will review recent developments in single cell sequencing technologies and their applications in cancer research. We will also explain how insights generated from single cell sequencing can be used to develop novel diagnostic and therapeutic approaches to conquer cancer.
- 20 September 2020
- Published by Springer Nature
- Vol. 1255, 109-121
Single Cell Sequencing: A New Dimension in Cancer Diagnosis and Treatment
Cancer is one of the leading causes of death worldwide and well known for its complexity. Cancer cells within the same tumor or from different tumors are highly heterogeneous. Furthermore, stromal and immune cells within tumor microenvironment interact with cancer cells to play important roles in how tumors progress and respond to different treatments. Recent advances in single cell technologies, especially massively parallel single cell sequencing, have made it possible to analyze cancer cells and cells in its tumor microenvironment in parallel with unprecedented high resolution. In this chapter, we will review recent developments in single cell sequencing technologies and their applications in cancer research. We will also explain how insights generated from single cell sequencing can be used to develop novel diagnostic and therapeutic approaches to conquer cancer.
- Open Access
Single-cell Analysis of ACE2 Expression in Human Kidneys and Bladders Reveals a Potential Route of 2019-nCoV Infection
Since December 2019, a novel coronavirus named 2019 coronavirus (2019-nCoV) has emerged in Wuhan of China and spread to several countries worldwide within just one month. Apart from fever and respiratory complications, acute kidney injury has been observed in some patients with 2019-nCoV. In a short period of time, angiotensin converting enzyme II (ACE2), have been proposed to serve as the receptor for the entry of 2019-nCoV, which is the same for severe acute respiratory syndrome coronavirus (SARS). To investigate the possible cause of kidney damage in 2019-nCoV patients, we used both published kidney and bladder cell atlas data and an independent unpublished kidney single cell RNA-Seq data generated in-house to evaluate ACE2 gene expressions in all cell types in healthy kidneys and bladders. Our results showed the enriched expression of all subtypes of proximal tubule cells of kidney and low but detectable levels of expression in bladder epithelial cells. These results indicated the urinary system is a potential route for 2019-nCoV infection, along with the respiratory system and digestion system. Our findings suggested the kidney abnormalities of SARS and 2019-nCoV patients may be due to proximal tubule cells damage and subsequent systematic inflammatory response induced kidney injury. Beyond that, laboratory tests of viruses and related indicators in urine may be needed in some special patients of 2019-nCoV.
Open AccessSingle-cell Analysis of ACE2 Expression in Human Kidneys and Bladders Reveals a Potential Route of 2019-nCoV Infection
Since December 2019, a novel coronavirus named 2019 coronavirus (2019-nCoV) has emerged in Wuhan of China and spread to several countries worldwide within just one month. Apart from fever and respiratory complications, acute kidney injury has been observed in some patients with 2019-nCoV. In a short period of time, angiotensin converting enzyme II (ACE2), have been proposed to serve as the receptor for the entry of 2019-nCoV, which is the same for severe acute respiratory syndrome coronavirus (SARS). To investigate the possible cause of kidney damage in 2019-nCoV patients, we used both published kidney and bladder cell atlas data and an independent unpublished kidney single cell RNA-Seq data generated in-house to evaluate ACE2 gene expressions in all cell types in healthy kidneys and bladders. Our results showed the enriched expression of all subtypes of proximal tubule cells of kidney and low but detectable levels of expression in bladder epithelial cells. These results indicated the urinary system is a potential route for 2019-nCoV infection, along with the respiratory system and digestion system. Our findings suggested the kidney abnormalities of SARS and 2019-nCoV patients may be due to proximal tubule cells damage and subsequent systematic inflammatory response induced kidney injury. Beyond that, laboratory tests of viruses and related indicators in urine may be needed in some special patients of 2019-nCoV.
- Top Cited PapersOpen Access
- 4 February 2020
- Published by Cold Spring Harbor Laboratory in bioRxiv
Specific ACE2 Expression in Cholangiocytes May Cause Liver Damage After 2019-nCoV Infection
A newly identified coronavirus, 2019-nCoV, has been posing significant threats to public health since December 2019. ACE2, the host cell receptor for severe acute respiratory syndrome coronavirus (SARS), has recently been demonstrated in mediating 2019-nCoV infection. Interestingly, besides the respiratory system, substantial proportion of SARS and 2019-nCoV patients showed signs of various degrees of liver damage, the mechanism and implication of which have not yet been determined. Here, we performed an unbiased evaluation of cell type specific expression of ACE2 in healthy liver tissues using single cell RNA-seq data of two independent cohorts, and identified specific expression in cholangiocytes. The results indicated that virus might directly bind to ACE2 positive cholangiocytes but not necessarily hepatocytes. This finding suggested the liver abnormalities of SARS and 2019-nCoV patients may not be due to hepatocyte damage, but cholangiocyte dysfunction and other causes such as drug induced and systemic inflammatory response induced liver injury. Our findings indicate that special care of liver dysfunction should be installed in treating 2019-nCoV patients during the hospitalization and shortly after cure.
- 4 February 2020
- Published by Cold Spring Harbor Laboratory in bioRxiv
Top Cited PapersOpen AccessSpecific ACE2 Expression in Cholangiocytes May Cause Liver Damage After 2019-nCoV Infection
A newly identified coronavirus, 2019-nCoV, has been posing significant threats to public health since December 2019. ACE2, the host cell receptor for severe acute respiratory syndrome coronavirus (SARS), has recently been demonstrated in mediating 2019-nCoV infection. Interestingly, besides the respiratory system, substantial proportion of SARS and 2019-nCoV patients showed signs of various degrees of liver damage, the mechanism and implication of which have not yet been determined. Here, we performed an unbiased evaluation of cell type specific expression of ACE2 in healthy liver tissues using single cell RNA-seq data of two independent cohorts, and identified specific expression in cholangiocytes. The results indicated that virus might directly bind to ACE2 positive cholangiocytes but not necessarily hepatocytes. This finding suggested the liver abnormalities of SARS and 2019-nCoV patients may not be due to hepatocyte damage, but cholangiocyte dysfunction and other causes such as drug induced and systemic inflammatory response induced liver injury. Our findings indicate that special care of liver dysfunction should be installed in treating 2019-nCoV patients during the hospitalization and shortly after cure.
- Open Access
- 23 October 2019
- Published by Oxford University Press (OUP) in Briefings in Bioinformatics
- Vol. 21 (5) , 1581-1595
Evaluation of single-cell classifiers for single-cell RNA sequencing data sets
Single-cell RNA sequencing (scRNA-seq) has been rapidly developing and widely applied in biological and medical research. Identification of cell types in scRNA-seq data sets is an essential step before in-depth investigations of their functional and pathological roles. However, the conventional workflow based on clustering and marker genes is not scalable for an increasingly large number of scRNA-seq data sets due to complicated procedures and manual annotation. Therefore, a number of tools have been developed recently to predict cell types in new data sets using reference data sets. These methods have not been generally adapted due to a lack of tool benchmarking and user guidance. In this article, we performed a comprehensive and impartial evaluation of nine classification software tools specifically designed for scRNA-seq data sets. Results showed that Seurat based on random forest, SingleR based on correlation analysis and CaSTLe based on XGBoost performed better than others. A simple ensemble voting of all tools can improve the predictive accuracy. Under nonideal situations, such as small-sized and class-imbalanced reference data sets, tools based on cluster-level similarities have superior performance. However, even with the function of assigning ‘unassigned’ labels, it is still challenging to catch novel cell types by solely using any of the single-cell classifiers. This article provides a guideline for researchers to select and apply suitable classification tools in their analysis workflows and sheds some lights on potential direction of future improvement on classification tools.
- 23 October 2019
- Published by Oxford University Press (OUP) in Briefings in Bioinformatics
- Vol. 21 (5) , 1581-1595
Open AccessEvaluation of single-cell classifiers for single-cell RNA sequencing data sets
Single-cell RNA sequencing (scRNA-seq) has been rapidly developing and widely applied in biological and medical research. Identification of cell types in scRNA-seq data sets is an essential step before in-depth investigations of their functional and pathological roles. However, the conventional workflow based on clustering and marker genes is not scalable for an increasingly large number of scRNA-seq data sets due to complicated procedures and manual annotation. Therefore, a number of tools have been developed recently to predict cell types in new data sets using reference data sets. These methods have not been generally adapted due to a lack of tool benchmarking and user guidance. In this article, we performed a comprehensive and impartial evaluation of nine classification software tools specifically designed for scRNA-seq data sets. Results showed that Seurat based on random forest, SingleR based on correlation analysis and CaSTLe based on XGBoost performed better than others. A simple ensemble voting of all tools can improve the predictive accuracy. Under nonideal situations, such as small-sized and class-imbalanced reference data sets, tools based on cluster-level similarities have superior performance. However, even with the function of assigning ‘unassigned’ labels, it is still challenging to catch novel cell types by solely using any of the single-cell classifiers. This article provides a guideline for researchers to select and apply suitable classification tools in their analysis workflows and sheds some lights on potential direction of future improvement on classification tools.
- Open Access
- 18 December 2017
- Published by Wiley in Journal of Neurochemistry
- Vol. 145 (1) , 68-79
Mechanisms of neuroprotection against ischemic insult by stress‐inducible phosphoprotein‐1/prion protein complex
Stress‐inducible phosphoprotein 1 (STI1) acts as a neuroprotective factor in the ischemic brain and its levels are increased following ischemia. Previous work has suggested that some of these STI1 actions in a stroke model depend on the recruitment of bone marrow‐derived stem cells to improve outcomes after ischemic insult. However, STI1 can directly increase neuroprotective signaling in neurons by engaging with the cellular prion protein (PrPC) and activating α7 nicotinic acetylcholine receptors (α7nAChR). Given that α7nAChR activation has also been involved in neuroprotection in stroke, it is possible that STI1 can have direct actions on neurons to prevent deleterious consequences of ischemic insults. Here, we tested this hypothesis by exposing primary neuronal cultures to 1‐h oxygen‐glucose deprivation (OGD) and reperfusion and assessing signaling pathways activated by STI1/PrPC. Our results demonstrated that STI1 treatment significantly decreased apoptosis and cell death in mouse neurons submitted to OGD in a manner that was dependent on PrPC and α7nAChR, but also on the activin A receptor 1 (ALK2), which has emerged as a signaling partner of STI1. Interestingly, pharmacological inhibition of the ALK2 receptor prevented neuroprotection by STI1, while activation of ALK2 receptors by bone morphogenetic protein 4 (BMP4) either before or after OGD was effective in decreasing neuronal death induced by ischemia. We conclude that PrPC/STI1 engagement and its subsequent downstream signaling cascades involving α7nAChR as well as the ALK2 receptor may be activated in neurons by increased levels of STI1. This signaling pathway protects neurons from ischemic insults. image
- 18 December 2017
- Published by Wiley in Journal of Neurochemistry
- Vol. 145 (1) , 68-79
Open AccessMechanisms of neuroprotection against ischemic insult by stress‐inducible phosphoprotein‐1/prion protein complex
Stress‐inducible phosphoprotein 1 (STI1) acts as a neuroprotective factor in the ischemic brain and its levels are increased following ischemia. Previous work has suggested that some of these STI1 actions in a stroke model depend on the recruitment of bone marrow‐derived stem cells to improve outcomes after ischemic insult. However, STI1 can directly increase neuroprotective signaling in neurons by engaging with the cellular prion protein (PrPC) and activating α7 nicotinic acetylcholine receptors (α7nAChR). Given that α7nAChR activation has also been involved in neuroprotection in stroke, it is possible that STI1 can have direct actions on neurons to prevent deleterious consequences of ischemic insults. Here, we tested this hypothesis by exposing primary neuronal cultures to 1‐h oxygen‐glucose deprivation (OGD) and reperfusion and assessing signaling pathways activated by STI1/PrPC. Our results demonstrated that STI1 treatment significantly decreased apoptosis and cell death in mouse neurons submitted to OGD in a manner that was dependent on PrPC and α7nAChR, but also on the activin A receptor 1 (ALK2), which has emerged as a signaling partner of STI1. Interestingly, pharmacological inhibition of the ALK2 receptor prevented neuroprotection by STI1, while activation of ALK2 receptors by bone morphogenetic protein 4 (BMP4) either before or after OGD was effective in decreasing neuronal death induced by ischemia. We conclude that PrPC/STI1 engagement and its subsequent downstream signaling cascades involving α7nAChR as well as the ALK2 receptor may be activated in neurons by increased levels of STI1. This signaling pathway protects neurons from ischemic insults. image