CellNetdb online database

What is CellNetdb

The cellular processes in a cell are achieved by groups of genes acting in concert to shape the cellular responses. Understanding how the gene network organized to mediate genotype-phenotype relationship is fundamental to global insights into the biological functions and disease etiology. CellNetdb is a compendium of cell-type-specific interactome networks across 44 human tumor types spanning over 2 million single cells.

Update

2023.10

Cancers

Cells

2.2M

Donors

563

Projects

Modules

Please refer to the following publication:

Zekun Li^, Gerui Liu^, Xiaoxiao Yang^, Meng Shu, Wen Jin, Yang Tong, Xiaochuan Liu, Yuting Wang, Jiapei Yuan*, Yang Yang* (2024) An atlas of cell-type-specific interactome networks across 44 human tumor types. Genome Medicine 16(1):30

Tips: Please wait for loading , and then select the Cancer Type and click the "Submit" button.

Cancer Type

Cell Type

Marker Gene

Tips: Please wait for loading , and then select the Cancer Type and click the "Submit" button.

Cancer Type

Cell Type

Marker Gene

Tips: Please wait for loading , and then select the Cell Type and click the "Submit" button.

Cell Type

Subtype

Marker Gene

Download

Gene prioritization

Gene prioritization

Help

Taxonomy

The taxonomy of cell types across different human cancers or pan-cancer infiltrating immune cells were displayed in this module. By selecting cancer/cell type and marker gene, users could visualize the cell-type taxonomy and gene expression pattern of cell-type maker genes. The cell-type taxonomy is displayed by the UMAP plot and the gene expression pattern is displayed by hexagon bin plot. Users could select the ‘Communication’ panel to view the integrated cell-cell communication. In addition, users could also obtain the number of single cells per cell type from the 'Summary' panel.

CellNet

Two levels of cell-type-specific networks are provided, including Cancer CellNet (cell-type-specific networks of different cell types across 44 cancer types) and TIME CellNet (cell-type-specific networks of different infiltrating immune cell types within the pan-cancer TIME). After selecting a cancer type or an immune cell type, users could query a specific gene in a cell type to obtain a subnetwork. The resulting subnetwork, which depicts the interactions between the queried gene and its neighboring genes within the cell-type-specific network, is presented in a visually dynamic network graph format.

CellNet: Edge

Users could obtain the information of edges and nodes in the cell-type-speicific network by selecting the 'Edge' panel. This panel provides a tabular view of the cell-type-specific network. It includes:

Cancer type: The selected cancer type.
Cell type: The selected cell type.
Gene 1: The queried gene.
Gene 2: The neighbor gene of the queried gene.
Weight: The weight of the network connection between Gene1 and Gene2.
Reference network: The reference network which SCINET used to reconstruct cell-type-specific networks.

CellNet: Expression

Users could visualize the gene expression patterns of all genes present in the queried subnetwork. It is displayed by dot-plot heatmaps where the color intensity represents the average expression level and the size of dot represents the percentage of cells within each cell type in the specific cancer type.

CellNet: Mutation

Users could browse or search somatic mutations in gene which are involved in the queried subnetwork by selecting the 'Mutation' panel. This section provides a tabular view of curated somatic mutations. It includes:

Gene: Gene involved in the queried subnetwork.
Mutation genome position: The genomic coordinates of the somatic mutation.
Mutation strand: The strand information of the somatic mutation.
Mutation AA: The amino acid substitution induced by the somatic mutation.
Mutation Description: The classification of the somatic mutation(substitution, deletion, insertion, complex, fusion, unknown etc.).
HGVSG: Human Genome Variation Society genomic syntax (3' shifted).

CellNet: GO

Users could browse or search GO terms significantly enriched for genes involved in the queried subnetworks by selecting the 'GO' panel. This panel provides a tabular view of enriched GO terms. It includes:

GO term ID: The ID of significantly enriched GO term.
GO term name: The name of significantly enriched GO term.
GeneRatio: The ratio of genes involved in the queried subnetwork which are annotated in a certain GO term.
BgRatio: The ratio of all expressed genes which are annotated in a certain GO term.
Fold enrichment: The quotient of “GeneRatio” and “Bgratio”.
Type: The type of GO terms (BP for Biological Process,MF for Molecular Function and CC for Cellular Component).
P-value: The P-value calculated by Fisher's exact test.
FDR: P-values of enrichment analysis were corrected using the Benjamini-Hochberg method.
Plot: Clicking the 'Plot' button to visualize the result of GO enrichment. A dynamic network graph is used to depict genes annotated in the GO term.

CellNet: Disease

Users could browse or search diseases significantly enriched for genes involved in the queried subnetworks by selecting the 'Disease' panel. This panel provides a tabular view of enriched diseases. It includes:

Disease: The disease name.
GeneRatio: The ratio of genes involved in the queried subnetwork which are annotated as genes associated with a certain disease.
BgRatio: The ratio of all expressed genes which are annotated as genes associated with a certain disease.
Fold enrichment: The quotient of “GeneRatio” and “Bgratio”.
P-value: The P-value calculated by Fisher's exact test.
FDR: P-values of enrichment analysis were corrected using the Benjamini-Hochberg method.
Plot: Clicking the 'Plot' button to visualize the result of disease enrichment.

CellNet: Survival

Users could browse or search genes involved in the queried subnetwork whose expression level are associated with patients' overall survival. The table includes:

Gene: The gene involved in the queried subnetwork.
Project: The project with overall survival data used for survival analysis.
Donors: The number of donors in the project.
P-value: The P-value calculated using log-rank test.
HR: Hazard ratio calculated using Cox proportional hazards regression.
Cox P-value: The P-value is calculated using Cox proportional hazards regression.
Plot: Clicking the 'Plot' button to visualize the Kaplan-Meier curve for survival analysis.

CellNet: Communication

Users could browse or search ligand-receptor pairs involved in the queried subnetwork by selecting the 'Communication' panel. This section provides a tabular view of cell-cell communication. It includes:

Ligand-receptor pair: The name of ligand-receptor pair.
Gene in network: The gene of subnetwork participated in cell-cell communication.
Ligand: The ligand in cell-cell communication.
Receptor: The receptor in cell-cell communication.
Communication score: The overall strength of cell-cell communication.
Plot: Clicking the 'Plot' button to visualize the result of cell-cell communication.

Analysis

Two levels of analyses are provided, including Cancer Analysis (based on cell-type-specific networks of different cell types across 44 cancer types) and TIME Analysis (based on cell-type-specific networks of different immune cell types within the pan-cancer TIME). Users could upload their files in this format:

GENE1

GENE2

GENE3

. . .

Analysis: Gene prioritization

Users could explore the prioritization of genes in cell-type-network based on genes that users uploaded by selecting the 'Gene prioritization' panel. The table includes:

Cell type: The cell type selected in analysis.
Subtype: The subtypes of selected cell type.
Number of seeds: The number of seed genes uploaded by users that appear in corresponding cell-type-specific network.
Prioritization: Clicking the ‘Show score' button to visualize prioritization of genes in corresponding cell-type-network.

About

What is CellNetdb

The cellular processes in a cell are achieved by groups of genes acting in concert to shape the cellular responses. Disruption of the human gene network could impair cellular functions and ultimately results in various diseases. Understanding how the gene network organized to mediate genotype-phenotype relationship is fundamental to global insights into the biological functions and disease etiology. Therefore, A reference gene network is needed to provide a scaffold for understanding biological mechanisms generally or within specific spatiotemporal context.

The CellNetdb is designed for users to explore cell-type-specific gene networks across various cancer types and pan-cancer tumor immune microenvironment (TIME). The networks were reconstructed by integrating a large compendium of curated scRNA-seq datasets using the SCINET method. It incoporates 44 different cancer types, involving 1,897,076 single cells from 36 solid tumor types and 310,965 single cells from 8 hematological malignancy types. Besides, pan-cancer cell atlas of infiltrating immune cells, including B cells, myeloid cells, CD4+ T cells, CD8+ T cells and B cells, are also integrated into the database. There are seven different fucntional modules as follows.

Taxonomy: The taxonomy of cell types across different cancer types or pan-cancer TIME.
Network: The cell-type-specific gene networks reconstructed for each cell type.
Somatic mutation: The somatic mutation which are located in genes involved in the cell-type-specific networks.
GO enrichment: The GO terms enriched for all genes involved in the network.
Disease enrichment: The diseases enriched for all genes involved in the network.
Survival: The relationship between genes involved in the network and patients' overall survival.
Cell-cell communication: The crosstalk between cell-type-specific network and intercellular network.

The Pipeline of Database Construction

Statistics of the CellNetdb Database

Acknowledgements

The CellNetdb was construted using R package shiny-server. R packages shinydashboard, shinythemes, dashboardthemes, shinyBS, shinyjs, shinybusy, shinycssloaders, plotly, networkD3 were used in interactive interface design. The R package DT which provides an R interface to the JavaScript library DataTables were applied to display tables.

How to cite

An atlas of cell-type-specific interactome networks across 44 human tumor types

Zekun Li^, Gerui Liu^, Xiaoxiao Yang^, Meng Shu, Wen Jin, Yang Tong, Xiaochuan Liu, Yuting Wang, Jiapei Yuan*, Yang Yang* (2024). Genome Medicine 16(1):30

Contact Information

Yang Yang (Corresponding Author)

Email: yy(AT)tmu.edu.cn

Gerui Liu (First Author)

Email:Liugerui(AT)tmu.edu.cn

Zekun Li (First Author)

Email:zeklee(AT)tmu.edu.cn

Affilation: Department of Bioinformatics, School of Basic Medical Sciences, Tianjin Medical University

Address: 22 Qixiangtai Road, Heping District, Tianjin 300070, China

Any comments or suggestion are welcome to improve CellNetdb.

Map







点标记