EMA (version 1.4.3)

genes.selection: Genes selection

Description

Selects the most variant genes of a expression dataset. For each gene, the difference between the quantile with probability (1 - probs) and the quantile with probability probs is computed. Default for probs value is set to 0.25 such that the difference corresponds to the Inter Quartile Range (IQR).

Usage

genes.selection(data, thres.diff, thres.num, probs=0.25)

Arguments

data

A expression matrix, genes on rows and samples on columns

thres.diff

Difference threshold - Genes are selected based on the difference threshold

thres.num

Genes number threshold - Selects the N genes with the highest difference value

probs

probability value used to compute both quantiles

Value

• The name of the selected genes.

Details

The difference is computed for each genes. If the thres.diff option is chosen, the most variant genes are selected according to the difference threshold. If the thres.num option is chosen, the genes are ordered according to their difference value level, and the N first genes are selected.

See Also

IQR, quantile

Examples

data(marty)
data.f<-expFilter(marty, graph=FALSE)

##Select the 50 most variant genes amongst the 1000 first genes
sel<-genes.selection(data.f[1:1000,], thres.num=50)