prcompPlot: PCA plot

In JasonR055/Harman: The removal of batch effects from datasets using a PCA and constrained optimisation based technique

Description

Generates a Principle Component plot for data.frames, matrices, or a pre-made prcomp object.

Usage

prcompPlot(object, pc_x = 1, pc_y = 2, scale = FALSE, colFactor = NULL,
  pchFactor = NULL, palette = "rainbow", legend = TRUE, ...)

Arguments

object	data.frame, matrix or prcomp object.
pc_x	integer, principle component for the plot x dimension.
pc_y	integer, principle component for the plot y dimension.
scale	logical, whether to scale to unit variance before PCA.
colFactor	factor or vector, colour the points by this factor, default is NULL.
pchFactor	factor or vector, point-type by this factor, default is NULL.
palette	string, the function to call to create a vector of contiguous colours with levels(colFactor) steps.
legend	logical, whether to display a legend on the plot.
...	further arguments passed to or from other methods.

Details

A data.frame object will be coerced internally to a matrix. Matrices must be of type double or integer. The prcompPlot function will then perform a principle component analysis on the data prior to plotting. The function is call is prcomp(t(object), retx=TRUE, center=TRUE, scale.=scale). Instead of specifying a data.frame or matrix, a pre-made prcomp object can be given to prcompPlot. In this case, care should be taken in setting the appropriate value of scale.. If a vector is given to colFactor or pchFactor, they will be coerced internally to factors.

For the default NULL values of colFactor and pchFactor, all colours will be black and circles the point type, respectively.

Value

None

See Also

prcomp rainbow

Examples

library(HarmanData)
data(IMR90)
expt <- imr90.info$Treatment
batch <- imr90.info$Batch
prcompPlot(imr90.data, colFactor=expt)
pca <- prcomp(t(imr90.data), scale.=TRUE)
prcompPlot(pca, 1, 3, colFactor=batch, pchFactor=expt, palette='topo.colors',
main='IMR90 PCA plot of Dim 1 and 3')

JasonR055/Harman documentation built on May 7, 2019, 10:32 a.m.