p.adjust: Improved p.adjust

In drmjc/mjcstats: MJC's general statistics library

Description

A new version of p.adjust which allows for qvalue estimation.

A replacement to the standard p.adjust.methods, including the Storey-Tibshirani positive FDR

Usage

  p.adjust(p, method = p.adjust.methods, n = length(p))

  p.adjust.methods

Arguments

p	a vector of p-values
method	see p.adjust.methods
n	optional number of observations

Format

chr [1:9] "holm" "hochberg" "hommel" "bonferroni" "BH" "BY" ...

Details

qvalue is known to print ERROR's if the input p-values are not suitably distributed. This function uses qvalue2, which tries to run qvalue with 2 different settings, else it falls back to using “BH” to estimate FDR.

Author(s)

Mark Cowley 2008-05-05

Mark Cowley

See Also

p.adjust p.adjust.methods qvalue2

p.adjust.methods p.adjust

Examples

pvals <- c(runif(100,0,0.05),runif(1e03,0,1))
bh    <- p.adjust(pvals, "BH")
bonf  <- p.adjust(pvals, "bonferroni")
qvals <- p.adjust(pvals, "qvalue")
## Not run: 
  par(mfrow=c(2,2))
  hist(pvals)
  hist(bh)
  hist(bonf)
  hist(qvals)

## End(Not run)

drmjc/mjcstats documentation built on May 15, 2019, 2:41 p.m.