While you might feel confused when you have no idea about dealing with cell communication analysis, it is because you do not possess a specific cellchat database of your own organism. Normally, this can only be done in humans or mice, or by matching homologous genes in your species. Now we provide a method that utilizes STRINGdb to predict interactions of your specific species as long as it has a reference genome. Alternatively, if you input a DGE list, we could assist you in predicting interactions and determining the reasons behind differential gene expression!
devtools::install_github("HallDaveNineteenNintysix/tocelldb")
If you have a list of genes that you interested in, such as a DGE list, you can use 'get_nodes' function to get predicted interactions from STRINGdb.
# Do not run
get_nodes(identifiers, # A character vector of genes, such as c("CDC42","CDK1",
# "KIF23","PLK1","RAC2","RACGAP1","RHOA","RHOB").
species = 9606, # NCBI taxon identifiers (e.g. Human is 9606, see: STRING organisms at
# https://string-db.org/cgi/input.pl?input_page_active_form=organisms/).
required_score = 400, # threshold of significance to include a interaction,
# a number between 0 and 1000 (default depends on the network).
partners=FALSE, # Whether provides the interactions between your set of proteins and all the other STRING proteins.
limit=10, # 10 by default while partners set to TRUE. the number of interaction partners retrieved per protein (most confident interactions come first)
network_type = "functional", # network type: functional (default), physical.
show_query_node_labels = 0, # when available use submitted names in the preferredName
# column when (0 or 1) (default:0)
add_nodes = 10, # adds a number of proteins with to the network based on their confidence score,
# larger number will return more interactions.
caller_identity = "Anonymous" # your identifier for STRINGdb.
)
Let's have a try with Rhinolophus ferrumequinum (greater horseshoe bat).
library(tocelldb)
rfq_nodes <- get_nodes(c("PPAT","GPR143","GAD1","ALDH5A1","GOT2","NAT8L","ASPA","FOLH1"),
species = 59479,
required_score = 400,
network_type = "functional",
show_query_node_labels = 0,
add_nodes = 10,
caller_identity = "Dave"
)
Then, we use 'rfq_nodes' as input with make to build a cellchatdb which can be used in CellChat analysis.
# Usage of make_customdb()
# Do not run!
make_customdb(interaction_nodes, # nodes file generated from get_nodes().
gene_info = NULL, # gene & protein infomations generated from down_gtf() if you need to transform protein_id to gene.
enrich = NULL, # A given pathway type, such as "KEGG：rfq00250", or "GO", as you wish, but better follow the CellChat db format.
annotation = NULL # A given interaction type, one of "Secreted Signaling", "ECM-Recptor", "Cell-Cell Contact", "Non-protein Signaling".)
# Start here
rfq_cellchatdb <- make_customdb(rfq_nodes, enrich = "KEGG：rfq00250", annotation = "Secreted Signaling")
We can download 'Protein(FASTA)' and 'Annotation features(GTF)' from NCBI. Then upload 'Protein(FASTA)' file, normally names 'protein.faa' to STRINGdb to add new organisms. After processing, we can search for our interested protein or pathways, then export .tsv files (...as short tabular text output) for input of make_customdb() from a STRINGdb network, and 'genomic.gtf' as gene_info parameter.
# Usage of make_customdb()
rfq_cellchatdb <- make_customdb("nodes.tsv", gene_info = 'genomic.gtf',enrich = NULL, annotation = NULL)