qvalue {webbioc}R Documentation

Estimate pFDR q-values from p-values

Description

Applies the q-value multiple testing procedure to a group of unadjusted p-values.

Usage

qvalue(p, lambda=seq(0,0.95,0.05), pi0.meth="smoother", fdr.level=NULL, robust=F)

Arguments

p a vector of p-values
fdr.level a level at which to control the FDR
lambda the value of the tuning parameter to estimate pi0
pi0.meth either "smoother" or "bootstrap"; the method for automatically choosing tuning parameter in the estimation of pi0, the proportion of true null hypotheses
robust an indicator of whether it is desired to make the estimate more robust for small p-values and a direct finite sample estimate of pFDR

Value

call gives the function call
pi0 an estimate of the proportion of null p-values
qvalues a vector of the estimated q-values (the main quantity of interest)
pvalues a vector of the original p-values
significant if fdr.level is specified, and indicator of whether the q-value fell below fdr.level (taking all such q-values to be significant controls FDR at level fdr.level)

Author(s)

John D. Storey http://faculty.washington.edu/~jstorey/

References

Storey JD. (2002) A direct approach to false discovery rates. Journal of the Royal Statistical Society, Series B, 64: 479-498.


[Package webbioc version 1.4.0 Index]