Show a three- two-dimensional plot of a prcomp object or a matrix, using different symbols and colors for groups of data
pca2d(
pca,
components = 1:2,
col = NULL,
title = NULL,
new = FALSE,
axes.color = "black",
bg = "white",
radius = 1,
group = NULL,
shape = NULL,
palette = NULL,
fancy = FALSE,
biplot = FALSE,
biplot.vars = 5,
legend = NULL,
show.scale = FALSE,
show.labels = FALSE,
labels.col = "black",
show.axes = TRUE,
show.axe.titles = TRUE,
axe.titles = NULL,
show.plane = TRUE,
show.shadows = FALSE,
show.centroids = FALSE,
show.group.labels = FALSE,
show.ellipses = FALSE,
ellipse.ci = 0.95,
...
)pca3d(
pca,
components = 1:3,
col = NULL,
title = NULL,
new = FALSE,
axes.color = "grey",
bg = "white",
radius = 1,
group = NULL,
shape = NULL,
palette = NULL,
fancy = FALSE,
biplot = FALSE,
biplot.vars = 5,
legend = NULL,
show.scale = FALSE,
show.labels = FALSE,
labels.col = "black",
show.axes = TRUE,
show.axe.titles = TRUE,
axe.titles = NULL,
show.plane = TRUE,
show.shadows = FALSE,
show.centroids = FALSE,
show.group.labels = FALSE,
show.shapes = TRUE,
show.ellipses = FALSE,
ellipse.ci = 0.95
)
Either a prcomp object or a matrix with at least three columns
Vector of length 3 (pca3d) or 2 (pca2d) containing the components to be shown
Either a single value or a vector of length equal to number of rows, containing color definitions for the plot points to be shown
Window title
Use TRUE to open a new window
Axis color This option has no effect in pca2d.
Background color
Scaling item for the size of points to be shown. In pca2d, this corresponds to the cex parameter.
either NULL or a factor of length equal to number of rows. Factor levels can be used to automatically generate symbols and colors for the points shown
Either a single value or a character vector describing the shapes to be used when drawing data points. Allowed shapes are: sphere, tetrahaedron and cube, and may be abbreviated. In pca2d, the parameter is passed directly on to the pch option of the points() function.
Specifies the color palette when colors are automatically assigned to the groups. See Details.
set show.labels, show.shadows, show.centroids and show.group.labels to TRUE.
Specify whether to show a biplot (see section ‘biplots’ below)
Specify which loading to show on the biplot (see section ‘biplots’ below)
TRUE for showing a numeric scale at the edges of the plot. This option has no effect in pca2d.
TRUE for showing labels (taken from the coordinate matrix or the prcomp object). Alternatively, a vector with labels to be shown next to the data points.
Single value or vector describing the colors of the labels.
TRUE to show the axes. This option has no effect in pca2d.
If TRUE, show axe titles (PC 1, PC 2 etc.) This option has no effect in pca2d.
A vector with two (pca2d) or three (pca3d) values containing the axe titles (corresponds to xlab and ylab in regular plot). If missing, but show.axe.titles is TRUE, axe titles will be generated automatically.
If TRUE, show a grey horizontal plane at y = 0. This option has no effect in pca2d.
If TRUE, show a "lollipop" representation of the points on the y = 0 plane: a vertical line joining the data point with the plane and a shadow. In pca2d, for each sample at (x,y), a grey line is drawn from (x,y) to (x,0).
If TRUE and the group variable is defined, show cluster centroids (using apropriate group symbols) and lines from each data point to the corresponding centroid.
Either TRUTH/FALSE or a vector equal to the number of unique values in
the group parameter. If set, labels for each of the defined
group will be shown at the group's centroid. If the value of the
parameter is TRUE, then the group names will be taken from the
group parameter. Otherwise, the values from this parameter will
be used.
A TRUTH/FALSE value indicating whether to show confidence interval ellipses or ellipsoids around each defined group
The confidence level of a pairwise confidence region for the CI. The default is 0.95, for a 95 the size of the ellipse being plotted.
For pca2d, any further argument will be passed on to the points() function.
A TRUTH/FALSE value indicating whether the different symbols (shapes) for the shown data points should be plotted (default TRUE).
Both pca2d and pca3d return invisibly a data frame which can be used to generate a legend for the figure. The data frame has as many rows as there are groups, and column with the group name, assigned color and assigned shape.
If option biplot is TRUE, a biplot showing both the PCA results
(samples) and variables is shown. This corresponds to the
biplot function which works for the prcomp
class objects. However, a biplot showing all variable loadings will be
unreadable if the data is highly dimensional (for example, gene
expression data). Therefore, the option biplot.vars specifies
which variables are shown on the biplot.
If biplot.vars is a vector of length larger than one, it will be
interpreted as a direct selection of the variables to be shown; for
example, for a prcomp object pca, the variable selection will
happen through pca$rotation[biplot.vars,].
If biplot.vars is a single number, then for each of the components shown, a number of variables equal to biplot.vars with the highest absolute loadings will be shown on the biplot.
The pca3d function shows a three dimensional representation of a PCA object or any other matrix. It uses the rgl package for rendering.
pca2d is the 2D counterpart. It creates a regular, two-dimensional plot on the standard graphic device. However, it takes exactly the same options as pca3d, such that it is easy to create 2D variants of the 3D graph.
Often, PCA visualisation requires using different symbols and colors for different groups of data. pca3d() and pca2d() aim at creating reasonable defaults, such that a simple call with two parameters -- the pca object and the vector with group assignments of the samples -- is sufficient for a basic diagnosis.
# NOT RUN {
data( metabo )
pca <- prcomp( metabo[,-1], scale.= TRUE )
pca3d( pca, group= metabo[,1] )
pca2d( pca, group= metabo[,1] )
## a bit more fancy:
## black background, white axes,
## centroids
pca3d( pca, group= metabo[,1],
fancy= TRUE, bg= "black",
axes.color= "white", new= TRUE )
# }
Run the code above in your browser using DataLab