multi_covar_pca: PCA plots of top two components colored by each covariate

In jdreyf/ezlimmaplot: Bioinformatics plots using limma and ggplot2, some from output of ezlimma

PCA plots of top two components colored by each covariate

Description

PCA plot of top two principal components using ggplot2 whose shape is defined by group and color in each plot by iterating over each covariate.

Usage

multi_covar_pca(
  object,
  pheno.df,
  name = "covar_pca",
  grp.var = "grp",
  covars = setdiff(colnames(pheno.df), grp.var),
  alpha = 1,
  all.size = NULL,
  facet = NULL,
  rm.leg.title = FALSE,
  labels = FALSE,
  manual.color = NULL,
  manual.shape = NULL,
  plot = TRUE
)

Arguments

object	Matrix-like object with features (e.g. genes) as rows and samples as columns.
pheno.df	Dataframe with rows as samples and columns as phenotypes.
name	Name of file to create. Set to NA to plot to screen instead of to file.
grp.var	Column of pheno.df to include as group variable in PCA plots.
covars	Columns of pheno.df to include as covariates in PCA plots.
alpha	Transparency, passed to geom_point.
all.size	Passed to geom_point size parameter to give size for all points without appearing in legend. ggplot2 default is size=2.
facet	A formula with columns in pheno.df to facet by.
rm.leg.title	Logical indicating if legend title should be removed.
labels	Logical, should sample labels be added next to points?
manual.color	Vector passed to scale_colour_manual for creating a discrete color scale. Vector length should be equal to number of levels in mapped variable.
manual.shape	Vector passed to scale_shape_manual for creating a discrete color scale. Vector length should be equal to number of levels in mapped variable.
plot	Logical; should plot be generated?

Details

PCA is calculated with prcomp.

Value

Invisibly, a list of the first two principal components appended to pheno.df for each covariate.

jdreyf/ezlimmaplot documentation built on Feb. 8, 2025, 2:25 a.m.