Protein-protein interaction analysis
test_PPI.Rd
test_PPI
search protein-protein interaction on significant candidate
based on a local STRING database.
Usage
test_PPI(
x,
contrasts = NULL,
species = "Human",
STRING.version = "11.5",
choose_scores = NULL,
score_cutoff = 400
)
Arguments
- x
A SummarizedExperiment/DEGdata output from
add_rejections
or a charachter vector containing candidate identifiers(SYMBOL, EntrezID, UniprotID or ENSEMBL).- contrasts
Character, specifies the contrasts to get significant candidates.
- species
The species name.
- STRING.version
Character(1),the version of STRING data.
- choose_scores
NULL or "combined_score" or a subset of c("neighborhood","fusion","cooccurence","coexpression", "experimental","database","textmining"). If is NULL or "combined_score", cutoff is on the scores combining above all evidence, otherwise is on the subset.
- score_cutoff
A numeric lager than 400, required lowest interaction scores. A score lowwer than 400 means the interaction is unconfident
Details
Identifiers in x are firstly transformed to entrezID and then mapped the relative STRING id.
Next the protein-protein interaction is exctracted from STRING.link.detail.
The local STRING data (including protein.info, protein.aliases and protein.links.detailed) is necessary for this function.
If it is absent, test_PPI
will download from STRING automatically according
species
and STRING.version
.
Examples
if (FALSE) {
# Load example
data(Silicosis_pg)
data <- Silicosis_pg
data_unique <- make_unique(data, "Gene.names", "Protein.IDs", delim = ";")
# Differential test
ecols <- grep("LFQ.", colnames(data_unique))
se <- make_se_parse(data_unique, ecols,mode = "delim")
filt <- filter_se(se, thr = 0, fraction = 0.4, filter_formula = ~ Reverse != "+" & Potential.contaminant!="+")
norm <- normalize_vsn(filt)
imputed <- impute(norm, fun = "MinProb", q = 0.05)
diff <- test_diff(imputed, type = "control", control = c("PBS"), fdr.type = "Storey's qvalue")
dep <- add_rejections(diff, alpha = 0.01,lfc = 2)
# PPI construct
if(check_PPI_depends()){
# Load STRING data. If local STRING is missing, PPI_res will download to system.file("PPIdata", "Mouse",package = "DEP2")
load_PPIdata(speciesname = "Mouse")
PPI_res <- test_PPI(dep, contrasts = "W4_vs_PBS", species = "Mouse",
STRING.version = "11.5",score_cutoff = 400)
head(PPI_res)
}
}