sigBiplot: Signature Biplot for principal compontents

In sidiropoulos/PSigA: A gene-signature ranking method based on sample density in PCA space

Description

Produces a bibplot of the selected signature's PCA. Based on ggbiplot and ggplot2.

Usage

sigBiplot(data, signature, groups = NULL, labels = NULL, pcs = c(1, 2),
  main = "", obs.size = 2, var.size = 3, var.scaled = FALSE,
  palette = "Paired", ...)

Arguments

data	gene expression matrix where rownames correspond to unique gene identifiers in signature format, and columns correspond to samples.
signature	character vector containing the signature's gene identifiers.
groups	optional factor variable indicating the groups that the observations belong to. If provided the points will be colored accordingly
labels	optional vector of labels for the observations
pcs	which PCs to plot. default: 1,2
main	plot title
obs.size	size of the points for the observations
var.size	size of the text for the variables
var.scaled	logical value. When set to TRUE the variable text size is proportional to the Euclidean Norm of the variable's PC loading; the variable's contribution to the variance.
palette	colorbrewer palette scheme to be used. Applicable only when groups are provided.
...	methods passed to ggbiplot.

Value

A ggplot2 plot

Examples

dummyData <- do.call(rbind, lapply(seq(0.1, 0.9, by = 0.1),
                     rnorm, n = 100, m = 6))
#add a row with bimodal gene expression
dummyData <- rbind(dummyData, c(rnorm(70, 6, 0.1), rnorm(30, 9, 0.1)))

rownames(dummyData) <- paste(rep("gene", nrow(dummyData)),
                             seq(1, nrow(dummyData)), sep = "")

dummySig <- c("gene1", "gene8", "gene9", "gene10")

sigBiplot(dummyData, dummySig)

groups <- factor(c(rep("GroupA", 70), rep("GroupB", 30)))
sigBiplot(dummyData, dummySig, groups, main = "dummySig")

sidiropoulos/PSigA documentation built on May 29, 2019, 9:58 p.m.