getMarkers: getMarkers
In mahmoudibrahim/genesorteR: Feature Ranking in Clustered Single Cell Data

Description Usage Arguments Value Author(s) See Also Examples

View source: R/getMarkers.R

getMarkers processes the output of sortGenes to select a relatively small set of marker genes.

1	getMarkers(gs, quant = 0.99, mutualInfo = FALSE, classEnt = FALSE)

`gs`	The output of `sortGenes()`.
`quant`	A number greater than zero and smaller than one. 0.99 by default. See Details.
`mutualInfo`	Logical. If `TRUE`, the mutual information between gene expression and cell clustering will be returned for each gene. `FALSE` by default.
`classEnt`	Logical. If `TRUE`, a "Cluster Shannon Index" is returned for each cluster. See Details. `FALSE` by default.

getMarkers returns a list with the following components:

`markers`	A character vector containing the names of selected marker features.
`maxScaledSpecScore`	A numeric vector of length `nrow(gs$specScore)`, including the maximum scaled specificity score for each gene across all cell clusters.
`gene_shannon_index`	A numeric vector of the same length as `maxScaledSpecScore`, including the Gene Shannon Index for each gene. See Details.
`mutInfo`	A numeric vector of the same length as `maxScaledSpecScore`, including the mutual information between the expression of each gene and the cell clustering.
`classEntropy`	A numeric vector of the same length as `ncol(specScore)`, including the Class Shannon Index for each cluster based on the selected marker genes.

Mahmoud M Ibrahim <mmibrahim@pm.me>

getPValues

#randomly generated data and cell clusters, almost no markers are found
set.seed(1234)
exp = matrix(sample(0:20,1000,replace=TRUE), ncol = 20)
rownames(exp) = sapply(1:50, function(x) paste0("g", x))
cellType = sample(c("cell type 1","cell type 2"),20,replace=TRUE)
sg = sortGenes(exp, cellType)
mm = getMarkers(sg,quant=0.95)
length(mm$markers) #only one marker gene was found


#"reasonably" separated clusters, with a few clear markers
data(sim)
gs = sortGenes(sim$exp, sim$cellType)
mm = getMarkers(gs,quant=0.95)
length(mm$markers)

#real data with three well separated clusters
data(kidneyTabulaMuris)
gs = sortGenes(kidneyTabulaMuris$exp, kidneyTabulaMuris$cellType)
mm = getMarkers(gs, quant = 0.99)
length(mm$markers) #we found 109 candidate markers
#we want to get a more focused list:
mm = getMarkers(gs, quant = 0.999)
length(mm$markers) #11 genes that can alone descriminate between the cell types