plotPvalCompa: Draw an XY plot comparing two series of p-values (or...

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

Description

Draw an XY plot comparing two series of p-values (or corrected p-values, or e-values).

Usage

plotPvalCompa(frame.to.plot, xlab = names(frame.to.plot)[1],
  ylab = names(frame.to.plot)[2], alpha = 0.05, score.name = "p-value",
  plot.result = TRUE, plot.colors = TRUE, plot.cex = 0.7,
  legend.cex = 0.9, legend.corner = "bottomright", ...)

Arguments

frame.to.plot	A data.frame comprizing two columns, with the respective p-values (corrected or not) to plot.
...	Additional parameters are passed to plot()
xlab=names(frame.to.plot)[1]	Passed to plot()
ylab=names(frame.to.plot)[2]	Passed to plot()
alpha=0.05	Alpha threshold to select positives and compute the contingency table.
score.name="p-value"	Score name to display on the legend (e.g. "p-value", "e-value", "fdr").
plot.result=TRUE	Draw a comparison plot
plot.colors=TRUE	Use colors for the plot
plot.cex=0.7	Point size for the plot (passed to graphics::plot() function).
legend.cex=0.9	Font size factor for legend. This allows to choose separate font sizes to draw the plot itself and the legend.
legend.corner="bottomright"	Position where legend should be placed.

Details

First version: 2015-03 Last modification: 2015-03

Author(s)

Jacques van Helden ([e-mail])

Examples

################################################################
## Load dataset from DenBoer, 2009 and define two groups of interest
library(denboer2009)
group1 <- "Bo"
group2 <- "Bt"

verbose(paste("Selecting samples for groups", group1, "and", group2), 1)
selected.groups <- c(group1, group2)
selected.samples <- denboer2009.pheno$sample.labels %in% selected.groups

## Selected expression profiles
selected.probesets <- apply(denboer2009.amp == "P", 1, sum) >= 30
expr <- denboer2009.expr[selected.probesets, selected.samples]
sample.labels <- denboer2009.pheno$sample.labels[selected.samples]
verbose(paste("Expression profiles: ", nrow(expr), "probesets x", ncol(expr), "samples"), 1)
table(sample.labels)

################################################################
## Compare p-values returned by Student and Welch t-tests
student.result <- tTestPerRow(x=expr,cl = sample.labels, var.equal = TRUE)
welch.result <- tTestPerRow(x=expr,cl = sample.labels, var.equal = FALSE)
plotPvalCompa(data.frame("Student.e.value"=student.result$table$e.value,
                         "Welch.e.value"=welch.result$table$e.value),
              score.name = "e-value",
              legend.corner="bottomright")

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