mutoss (version 0.1-3)

Qvalue: Storey's (2001) q-value Procedure...

Description

Storey's (2001) q-value Procedure

Usage

Qvalue(pValues, lambda=seq(0, 0.9, 0.05), pi0.method="smoother",
    fdr.level=NULL, robust=FALSE, smooth.df=3, smooth.log.pi0=FALSE,
    silent=FALSE)

Arguments

pValues

pValues to be used (only necessary input)

lambda

Value of the tuning parameter to be used

pi0.method

Method for automatically choosing tuning parameter in the estimation of pi_0. Either 'smoother' or 'bootstrap'

fdr.level

Level at which to control the FDR

robust

Logical, whether to make estimate more robust for small p-values.

smooth.df

Number of degrees of freedom to use when estimating pi_0 with the smoother.

smooth.log.pi0

Logical, if TRUE and pi0.method = 'smoother', pi0 will be estimated by applying a smoother to a scatterplot of log(pi_0) estimates against the tuning parameter lambda.

silent

logical scalar. If TRUE no output is generated.

Value

• A list containing:
• qValues
  A vector of the estimated q-values
• pi0
  An estimate of the proportion of null hypotheses
• errorControl
  A Mutoss S4 class of type errorControl, containing the type of error controlled by the function.

Details

The Qvalue procedure estimates the q-values for a given set of p-values. The q-value of a test measures the proportion of false positive incurred when that particular test is called sigificant. It gives the scientist a hypothesis testing error measure for each observed statistic with respect to the pFDR.

Note: If no options are selected, then the method used to estimate pi0 is the smoother method desribed in Storey and Tibshirani (2003). The bootstrap method is described in Storey, Taylor and Siegmund (2004).

References

Storey, John (2001). The Positive False Discovery Rate: A Baysian Interpretation and the Q-Value. The Annals of Statistics, Vol. 31, No. 6, 2013-2035.

Examples

pval <- c(runif(50), runif(50, 0, 0.01))
result <- Qvalue(pval)
result <- Qvalue(pval, lambda=0.5)