DEDS (version 1.46.0)

pairs-methods: Pairs Plot for DEDS Objects

Description

The function pairs-DEDS produces pairs plots of statistics or p values for DEDS-class objects.

Usage

"pairs"(x, subset=c(1:nrow(x$stats)), labels =
colnames(x$stats[,-1]), logit = FALSE,
diagonal = c("qqnorm", "boxplot", "density", "histogram", "none"),
lower = c("cor", "none"), groups.by.deds = TRUE, thresh = 0.05, reg.line
= NULL, smooth = FALSE, line.by.group = FALSE, diag.by.group = TRUE, lower.by.group =
FALSE, col = palette(), pch = 1:n.groups, lwd = 1, legend.plot =
length(levels(groups)) > 1, ...)

Arguments

x

An object of DEDS.

subset

A numeric vector indicating the subset of points to be plotted.

labels

A character vector specifying the names of the variables.

logit

A logical variable, if TRUE the variables are logged, useful when plotting p values.

diagonal

A character string specifying the type of plot to be applied in the diagonal panels.

diagonal="qqnorm":

qqnorm on the diagonal

diagonal="boxplot":

boxplot on the diagonal

diagonal="density":

density on the diagonal

diagonal="histogram":

hist on the diagonal

lower

A character string specifying the function to be applied in the lower panels.

lower="cor":

absolute correlation will be put on the lower panel

groups.by.deds

A logcial variable, if TRUE, points will be separated into groups according to their magnitude of q- or p-values by DEDS.

thresh

A numeric variable, if thresh<1, it="" specifies="" the="" threshold="" of="" significance="" in="" differential="" expression="" (de)="" for="" q-="" or="" p-values="" deds="" object;="" default is="" set="" at="" 0.05.="" if="" thresh>1, it specifies the number of top DE genes to be highlighted.

reg.line

A function name specifying the type of regression line to be plotted in the scatter plots. If reg.line=lm, linear regression line will be plotted; If reg.line=NULL, no regression line will be plotted in the scatter plot.

smooth

A logical variable specifying if smooth regression lines will be plotted in the scatter plots. If smooth=TRUE, a lowess line will be applied.

line.by.group

A logical variable specifying if the regression lines should be applied within groups.

diag.by.group

A logical variable specifying if the plot in the diagonal panels would be applied groupwise.

lower.by.group

A logical variable, if lower.by.group=TRUE and lower="cor", correlation coefficients will be calculated and printed separated according to groups in the lower panels.

col

A specification for the colors to be used for plotting different groups, see par.

pch

A specification for the type of points to be used for plotting different groups, see par.

lwd

A specification for the width of lines to be used if lines are plotted; see par.

legend.plot

A logical variable specifying if the legend will be plotted.

...

Extra parameters for plotting.

Details

The function pairs.DEDS implements a S3 method of pairs for DEDS. The DEDS class is a simple list-based class to store DEDS results and it is usually created by functions deds.pval, deds.stat, deds.stat.linkC. The list contains a "stat" component, which stores statistics or p values from various statistical tests. The function pairs.DEDS extracts the "stat" component and produces a matrix of scatterplot.

pairs.DEDS as a default highlights points (corresponding to genes) with adjusted p- or q-values less than a user defined threshold. The user can select among a series of options a plot for the diagonal panel; as a default, it produces a qqnorm for each column in the "stat" matrix. Both the diagonal and lower panels can be stratified by specifying the diag.by.group or lower.by.group arguments.

See Also

deds.stat, deds.pval, deds.stat.linkC, hist.DEDS, qqnorm.DEDS

Examples

X <- matrix(rnorm(1000,0,0.5), nc=10)
L <- rep(0:1,c(5,5))

# genes 1-10 are differentially expressed
X[1:10,6:10]<-X[1:10,6:10]+1
# DEDS
d <- deds.stat.linkC(X, L, B=200)

# pairs plot
pairs(d)
# plot regression line
pairs(d, reg.line=lm, lwd=2)
# histogram in the diagonal panel
pairs(d, diagonal="hist")
# boxplot on the diagonal panel and stratified
pairs(d, diagonal="boxplot", diag.by.group=TRUE)