Plot PCA For any assay within a SummarizedExperiment

Source: R/plotAssayPCA.R

Plot PCA for any assay within a SummarizedExperiment object

plotAssayPCA(x, ...)

# S4 method for class 'SummarizedExperiment'
plotAssayPCA(
  x,
  assay = "counts",
  colour = NULL,
  shape = NULL,
  size = NULL,
  label = "colnames",
  show_points = TRUE,
  pc_x = 1,
  pc_y = 2,
  trans = NULL,
  n_max = Inf,
  tol = sqrt(.Machine$double.eps),
  rank = NULL,
  ...
)

Arguments

x

An object containing an assay slot

...

Passed to geom_text and geom_point

assay

The assay to perform PCA on

colour, size

The column names to be used for colours and point/label size. Can be fixed values (e.g. size = 3) and can also be a manipulation of a column, e.g. colour = log10(totals)

shape

The column name(s) to be used for determining the shape or size of points. Can also be a fixed value

label

The column name to be used for labels. Will default to the column names of the SummarizedExperiment

show_points

logical(1). Display the points. If TRUE any labels will repel. If FALSE, labels will appear at the exact points

pc_x

numeric(1) The PC to plot on the x-axis

pc_y

numeric(1) The PC to plot on the y-axis

trans

character(1). Any transformative function to be applied to the data before performing the PCA, e.g. trans = "log2"

n_max

Subsample the data to this many points before performing PCA

tol, rank

Passed to prcomp

Value

A ggplot2 object

Details

Uses ggplot2 to create a PCA plot for the selected assay. Any numerical transformation prior to performing the PCA can be specified using the trans argument

Examples

data("se")
se$treatment <- c("E2", "E2", "E2", "E2DHT", "E2DHT", "E2DHT")
se$sample <- colnames(se)
plotAssayPCA(se, trans = "log1p", colour = "treatment", label = "sample")

plotAssayPCA(
  se, trans = "log1p", colour = "treatment", label = "sample",
  size = log10(totals), shape = 17
)

plotAssayPCA(
  se, trans = "log1p", colour = "treatment", label = "sample",
  show_points = FALSE
)