tTestPerRow.bootstrap.hist: Plot an histogram with a bootstrap test result.

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

Description

Plot an histogram from the result of tTestPerRow.bootstrap(). The histogram indicates the number of features (ordinate) with a given support (abcissa), i.e. the number of features declared significant in exactly X bootstrap iterations.

Usage

tTestPerRow.bootstrap.hist(bootstrap.result, xlab = "Support",
  ylab = "Number of features", support.statistics = "fdr",
  support.quantile = bootstrap.result$support.quantile, main = paste(sep =
  "", "Bootstrap support (", support.statistics, "<=", bootstrap.result$alpha,
  ")"), plot.dcdf = FALSE, plot.legend = TRUE, legend.corner = "topright",
  lwd = 1, ylim = NULL, ...)

Arguments

bootstrap.result	An object returned by the function tTestPerRow.bootstrap().
plot.legend	= TRUE Only valid if plot.dcdf is TRUE. Plot a legend indicating the number of features supported by a given number of the iterations.
...	Additional parameters are passed to hist().
support.statistics="fdr"	Statistics for which the support has to be displayed. Supported values: "fdr", "p.value", "e.value".
support.quantile=0.75	Minimal percent of support required to declare a feature positive. The default is to select features supported in 75% of bootstrap iterations.
xlab="Support"	Label for the X axis.
ylab="Number	of features" Label for the Y axis.
main=paste(sep="",	"Bootstrap support (", support.statistics,"<=", bootstrap.result$alpha, ")")
plot.dcdf=FALSE	Plot the decreasing CDF over the histogram, in "step" mode. This curve indicates the number of features supported by at least X bootstraps.
legend.corner="topright"	Corner to plo the legend.
lwd=1	Line width for the dcdf and support threshold lines.

Details

First version: 2015-04 Last modification: 2015-04

Author(s)

Jacques van Helden ([e-mail])

Examples

## First run the examples of tTestPerRow.bootstrap(), in order to get some result to plot
example("tTestPerRow.bootstrap")

## Plot the histogram
tTestPerRow.bootstrap.hist(student.bootstrap, col="#BBBBBB")

## Plot the histogram + the dcdf. 
tTestPerRow.bootstrap.hist(student.bootstrap, col="#BBBBBB", 
                           plot.dcdf=TRUE, lwd=2)

## Plot the histogram + the dcdf for the e-value. 
tTestPerRow.bootstrap.hist(student.bootstrap, support.statistics="e.value",
                           col="#BBBBBB", 
                           plot.dcdf=TRUE, lwd=2)

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