tmodAUC: Calculate AUC

Description

Calculate AUC

Usage

tmodAUC(
  l,
  ranks,
  modules = NULL,
  stat = "AUC",
  recalculate.ranks = TRUE,
  filter = FALSE,
  mset = "all"
)

Value

A matrix with the same number of columns as "ranks" and as many rows as there were modules to be tested.

Arguments

l: List of gene names corresponding to rows from the ranks matrix
ranks: a matrix with ranks, where columns correspond to samples and rows to genes from the l list
modules: optional list of modules for which to make the test
stat: Which statistics to generate. Default: AUC
recalculate.ranks: Filtering and removing duplicates will also remove ranks, so that they should be recalculated. Use FALSE if you don't want this behavior. If unsure, stay with TRUE
filter: Remove gene names which have no module assignments
mset: Which module set to use. "LI", "DC" or "all" (default: "all")

Details

tmodAUC calculates the AUC and U statistics. The main purpose of this function is the use in randomization tests. While tmodCERNOtest and tmodUtest both calculate, for each module, the enrichment in a single sorted list of genes, tmodAUC takes any number of such sorted lists. Or, actually, sortings -- vectors with ranks of the genes in each replicate.

Note that the input for this function is different from tmodUtest and related functions: the ordering of l and the matrix ranks does not matter, as long as the matrix ranks contains the actual rankings. Each column in the ranks matrix is treated as a separate sample.

Also, the `nodups` parameter which is available (and TRUE by default) for other tests cannot be used here. This means that the AUCs calculated here might be slightly different from the AUCs calculated with default parameters in tests such as the [tmodCERNOtest()]. Use `nodups=FALSE` with [tmodCERNOtest()] to obtain identical results as with `tmodAUC`.

Examples

Run this code

data(tmod)
l <- tmod_ids(tmod)
ranks <- 1:length(l)
res <- tmodAUC(l, ranks)
head(res)