logpadjust

This package is currently meant to provide access to the function logp.adjust, which is a modified version of stats::p.adjust that accepts log-space p-values and returns adjusted p-values in log-space.

While p.adjust could be used in the majority of cases by converting values to linear space, certain methods may generate p-values that exceed the minimum value (~1e-300). This package was written to provide a function that would enable multiple test correction methods on these values without losing information.

Installation

You can install the development version of logpadjust from GitHub with:

# install.packages("devtools")
devtools::install_github("andvon/logpadjust")

Example

The logp.adjust function should be able to be called and used in the same manner as stats::p.adjust.

library(logpadjust)
log_p_values <- log(c(0.005,0.1))
logp.adjust(p = log_p_values,method = "bonferroni")

The methods (which should match stats::p.adjust.methods) can be viewed with logp.adjust.methods.

If p-values were calculated in a space other than the natural log, specifying base will update the calculations.

log10_p_values <- log10(c(0.005,0.1))
logp.adjust(p = log10_p_values,method = "bonferroni", base = 10)

For additional help, please see the package's documentation.