Plot pretty PCA plots using custom ggplot theme.

Run PCA on the given data matrix and generates PC plots for the specified principal components.

plot_pca(
  X,
  pca_obj,
  npcs,
  pcs,
  color = NULL,
  color_upper = NULL,
  color_label = "",
  color_upper_label = "",
  color_scheme = NULL,
  color_scheme_upper = NULL,
  point_size = 0.5,
  point_alpha = 1,
  subsample = 1,
  show_var = TRUE,
  center = TRUE,
  scale = FALSE,
  title = "",
  show_plot = FALSE,
  ...
)

Arguments

X

Data matrix or data.frame on which to perform PCA. Must specify either X or pca_obj.

pca_obj

Output of previous run of plot_pca() to avoid re-computing SVDs (i.e., the PC loadings and scores) again. Must specify either X or pca_obj. Ignored if X is provided.

npcs

Number of top PCs to plot. Must specify either npcs or pcs.

pcs

Vector of which PCs to show. Must specify either npcs or pcs. Ignored if npcs is provided.

color

(Optional) Data vector to use as colors for lower ggplot panels.

color_upper

(Optional) Data vector to use as colors for upper ggplot panels.

color_label

Character string. Label for color legend title (used in lower ggplot panels).

color_upper_label

Character string for color_upper legend title (used in upper ggplot panels).

color_scheme

(Optional) Vector of colors to set manual color scheme corresponding to color_lower argument (i.e., the color scheme in the lower panels). If NULL (default), viridis color scheme is used.

color_scheme_upper

(Optional) Vector of colors to set manual color scheme corresponding to color_upper argument (i.e., the color scheme in the upper panels). If NULL (default), viridis color scheme is used.

point_size

Point size for [ggplot2::geom_point()].

point_alpha

Alpha value for [ggplot2::geom_point()].

subsample

Proportion of rows to subsample and plot.

show_var

Logical. Whether or not to show the proportion of variance explained in axes labels.

center

Logical. Whether or not to center data for PCA.

scale

Logical. Whether or not to scale data for PCA.

title

Character string. Title of plot.

show_plot

Logical. Should this plot be printed? Default FALSE.

...

Other arguments to pass to vthemes::theme_vmodern() or theme_function()

Value

A list of four:

plot

A ggplot object of the PC pair plots.

scores

A matrix with the PC scores.

loadings

A matrix with the PC loadings.

var.explained

A vector of the proportions of variance explained.

Examples

out <- plot_pca(X = iris[, -5], npcs = 3, color = iris$Species)
out$plot

iris2 <- data.frame(iris, z = rep(letters[1:2], length.out = nrow(iris)))
out <- plot_pca(X = iris2[, -c(5, 6)], npcs = 3,
               color = iris2$Species, color_upper = as.factor(iris2$z))
out$plot