fdc {arrayMagic}R Documentation

FDC (false discovery count)

Description

Estimate the FDC (false discovery count) through permutations

Usage

fdc(x, fac,
  teststatfun = "rowFtests",
  nrperm      = 100,
  nrgenesel   = c(10, 20, 40, 60, 80, 100, 200),
  ...)

Arguments

x Matrix.
fac Factor, with length(fac)=ncol(x).
teststatfun Character. Name of a function that takes arguments x and fac, and returns a list with component statistic. See for example rowFtests.
nrperm Numeric. Number of permutations.
nrgenesel Numeric. A vector with the 'number of genes' for which the FDC is to be calculated.
... Further arguments passed to code{teststatfun}.

Value

A list with elements stat: the test statistics; mpstat: median permuted test statistics; fdc: estimated false discovery counts; thresh: the threshholds associated with nrgenesel; nrgenesel

Author(s)

Wolfgang Huber <w.huber@dkfz.de>

See Also

rowFtests

Examples

   ## data matrix: 2000 genes, 16 samples
   x  <- matrix(runif(2000*16), ncol=16)
   ## 8 blue and 8 red samples
   fac <- factor(c(rep("blue", 8), rep("red", 8)))
   ## implant differential signal into the first 50 genes
   x[1:50, fac=="blue"] <- x[1:50, fac=="blue"] + 1

   res <- fdc(x, fac)
   plot(res$nrgenesel, res$fdc, pch=16, col="blue",
       xlab="Number of genes selected",
       ylab="Expected number of false discoveries")
   abline(a=0, b=1, col="red", lwd=2)

   qqplot(res$stat, res$mpstat, pch=".")
   abline(a=0, b=1, col="red", lwd=2)


[Package Contents]