mulitpleTestingCorrections.plotCorrectedVsPval: Plot corrected p-values versus nominal p-value

In jvanheld/stats4bioinfo: Utilities for the book "Statistics for bioinformatics"

Description

Plot the different multi-testing corrected statistics as a function of the nominal P-value.

Usage

mulitpleTestingCorrections.plotCorrectedVsPval(multitest.result,
  main = "Multitesting corrections", xlab = "p-value",
  ylab = "Multi-testing corrected statistics", alpha = 0.05,
  legend.corner = "topleft", legend.cex = 1, plot.legend = TRUE,
  plot.pch = c(p.value = 2, fdr = 4, qval.0 = 3, e.value = 1, fwer = 20),
  plot.col = c(p.value = "#000000", fdr = "#888888", qval.0 = "#666666",
  e.value = "#BBBBBB", fwer = "#444444"), plot.elements = c("p.value", "fdr",
  "qval.0", "e.value", "fwer"), ...)

Arguments

multitest.result	the list returned by the function multipleTestingCorrections().
...	Additional parameters are passed to plot()
main='Multitesting	corrections' main title of the plot
alpha=0.05	Threshold of significance (alpha).
plot.pch=c(p.value=2, fdr=4, qval.0=3, e.value=1, fwer=20)	Specific characters to distinguish the plotted statistics.
plot.col=c(p.value='#000000', fdr='#888888', qval.0='#666666', e.value='#BBBBBB', fwer='#444444')	Specific colors or gray levels to distinguish the plotted statistics.
plot.elements=c("p.value", "e.value", "fwer", "fdr", "qval.0")	Selection of elements to display on the plot.
plot.legend=TRUE	Plot a legend indicating the number of features declared significant with the alpha threshold on the selected statistics.
legend.corner="topleft"	corner wher the legend has to be placed.
legend.cex=1	Font size for the legend.
xlab="p-value"	Label for the X axis
ylab="Multi-testing	corrected statistics" Label for the Y axis

Author(s)

Jacques van Helden ([e-mail])

Examples

## To obtain the input list (multitest.result), run the example of
## stats4bioinfo::multipleTestingCorrections().
example(multipleTestingCorrections)

## Plot all the multiple testing corrections at once
mulitpleTestingCorrections.plotCorrectedVsPval(multitest.result)

## Compare e-value and FWER
mulitpleTestingCorrections.plotCorrectedVsPval(multitest.result, plot.elements=c("e.value","fwer"))

## Compare e-value and FDR. 
## This plot highlights the non-linear relationship between FDR and p-value.
mulitpleTestingCorrections.plotCorrectedVsPval(multitest.result, plot.elements=c("e.value","fdr"))

## Compare Benjamini-Hochberg (qval.0) and Storey-Tibshirani (fdr) estimates of FDR.
mulitpleTestingCorrections.plotCorrectedVsPval(multitest.result, plot.elements=c("qval.0","fdr"))

jvanheld/stats4bioinfo documentation built on May 20, 2019, 5:16 a.m.