Run PCA and generate plots — doPCA • einprot

Apply PCA to the assay defined by assayName and extract the top ncomponent components. For each pair defined in plotpairs, generate a collection of plots.

doPCA(
  sce,
  assayName,
  ncomponents = 10,
  ntop = Inf,
  plotpairs = list(c(1, 2)),
  maxNGroups = 10,
  maxTextWidthBarplot = NULL,
  colourBy = "group",
  subset_row = NULL,
  scale = FALSE
)

Arguments

sce: A SingleCellExperiment object.
assayName: Character scalar defining the assay in sce to use for the PCA.
ncomponents: Numeric scalar, the (maximal) number of components to extract. The actual number can be lower if the number of samples is too small.
ntop: Number of features (with highest variance) to use to generate the PCA. Will be passed on to scater::runPCA.
plotpairs: A list of numeric vectors of length 2, indicating which pairs of PCs to generate plots for.
maxNGroups: Numeric scalar, the maximum number of groups to display in the legend in the scatter plot in the combined plot. If there are more than maxNGroups groups, the legend is suppressed.
maxTextWidthBarplot: Numeric scalar giving the maximum allowed width for text labels in the bar plot of log-fold changes. If not NULL, the size of the labels will be scaled down in an attempt to keep the labels inside the canvas. Typically set to half the width of the plot device (in inches).
colourBy: Character scalar indicating the name of the column of colData(sce) that will be used to colour the points.
subset_row: Vector specifying the subset of features to use for dimensionality reduction. Can be a character vector of row names, an integer vector of row indices or a logical vector. Will be passed to scater::runPCA.
scale: Logical scalar indicating whether the values should be scaled before the PCA is applied. Will be passed to scater::runPCA.

Value

A list with the following components:

sce - the input sce, expanded with the calculated PCs, in addition the feature coefficients will be added to the rowData.
plotcoord - a list of ggplot objects containing coordinate plots for the desired pairs of components.
plotcombined - a list of ggplot objects containing combined coordinate, scree and coefficient plots for the desired pairs of components.
plotpairs - a ggpairs plot with all extracted components.

Author

Charlotte Soneson

Examples

sce <- readRDS(system.file("extdata", "mq_example", "1356_sce.rds",
                           package = "einprot"))
pca <- doPCA(sce, assayName = "log2_LFQ.intensity", ncomponents = 3)
pca$plotcoord$PC1_2

pca$plotcombined$PC1_2

pca$plotpairs