Cell-cell communications via secreting and receiving ligands frequently occur in multicellular organisms, which is a vital feature involving numerous biological processes.[1] Recent advancements in single-cell RNA sequencing (scRNA-seq) have effectively resolving cellular phenotype heterogeneity and cell-type composition of complex tissues, which enables systematic investigation of cell-cell communications at a single-cell resolution. However, the common practice to study chemical signal-dependent cell-cell communications with scRNA-seq relies heavily on the prior knowledge of ligand-receptor (LR) interaction pairs. Here, we introduce CellTalkDB, a comprehensive database of LR interaction pairs in human and mouse by text mining and manual verification of known protein-protein interactions (PPIs) in STRING.

[1] Xin Shao et al. New Avenues for Systematically Inferring Cell-Cell Communication: Through Single-Cell Transcriptomics Data.Protein Cell.2020 May 21.doi: 10.1007/s13238-020-00727-5

Shao et al., CellTalkDB: a manually curated database of ligand–receptor interactions in humans and mice, Briefings in Bioinformatics, 04 November 2020 (online), bbaa269, https://doi.org/10.1093/bib/bbaa269

Users can download the LR pairs in CellTalkDB and replace the underlying database in SoptSC, SingleCellSignalR and CellPhoneDB, etc. to identify significantly enriched LR pairs and to infer cell-cell communications.

To help users use CellTalkDB conveniently, we made the revised R package by only replacing the underlying database in SingleCellSignalR and keeping the other functions unchanged to guide users on how to use data on LR pairs in CellTalkDB for the analysis of cell-cell communications with transcriptomics data. Source package of scsrctdb-1.0 can be downloaded in the release page.

# download the source package of scsrctdb-1.0.tar.gz and install it
# ensure the right directory for scsrctdb-1.0.tar.gz
# ensure the dependency packages have been installed.('BiocManager','circlize','limma','igraph','gplots','grDevices','edgeR','SIMLR','data.table','pheatmap','stats','Rtsne','graphics','stringr','foreach','multtest','scran')
install.packages(pkgs = 'scsrctdb-1.0.tar.gz',repos = NULL, type = "source")

• For human scRNA-seq datasets

# based on 3,398 human LR pairs
library(scsrctdb)
cell_signal <- cell_signaling(data = data,
                              genes = genes,
                              cluster = cluster,
                              gene_resive = T,
                              species = 'homo sapiens')

• For mouse scRNA-seq datasets

# based on 2,033 mouse LR pairs
library(scsrctdb)
cell_signal <- cell_signaling(data = data,
                              genes = genes,
                              cluster = cluster,
                              gene_resive = T,
                              species = 'mus musculus')

• Use visualize() for plotting

visualize(cell_signal)

visualize(cell_signal,show.in = 1)

Note: we have added an extra parametergene_resiveto revise gene symbols according to NCBI Gene database (updated in April 28,2020) as CellTalkDB has been revised with it. For more information about how to use SingleCellSignalR, please refer to wiki page or SCA-IRCM/SingleCellSignalR

• LR pairs for human and mouse can be download indatabase/
• Annotation data for LR pairs can be downloaded indata/
• Raw data for reproduction of our results can be downloaded in the release page.

CellTalkDB repository was developed by Xin Shao. Should you have any questions, please contact Xin Shao at [email protected]. For more information, please refer to our published paper 10.1093/bib/bbaa269 or visit our website tcm.zju.edu.cn/celltalkdb