plotTangentSpace: Plot specimens in tangent space

Description

Function plots a set of Procrustes shape variables in tangent space along their principal component axes

Usage

plotTangentSpace(A, axis1 = 1, axis2 = 2, warpgrids = TRUE, mesh = NULL,
  label = NULL, groups = NULL, legend = FALSE, ...)

Arguments

A 3D array (p x k x n) containing Procrustes shape variables for a set of specimens

axis1

A value indicating which PC axis should be displayed as the X-axis (default = PC1)

axis2

A value indicating which PC axis should be displayed as the Y-axis (default = PC2)

warpgrids

A logical value indicating whether deformation grids for shapes along X-axis should be displayed

mesh

A mesh3d object to be warped to represent shape deformation along X-axis (when warpgrids=TRUE) as described in plotRefToTarget.

label

An optional vector indicating labels for each specimen are to be displayed (or if TRUE, numerical addresses are given)

groups

An optional factor vector specifying group identity for each specimen (see example)

legend

A logical value for whether to add a legend to the plot (only when groups are assigned).

...

Arguments passed on to prcomp. By default, plotTangentSpace will attempt to remove redundant axes (eigen values effectively 0). To override this, adjust the argument, tol, from prcomp.

Value

If user assigns function to object, returned is a list of the following components:

pc.summary

A table summarizing the percent variation explained by each pc axis, equivalent to summary of prcomp.

pc.scores

The set of principal component scores for all specimens.

pc.shapes

A list with the shape coordinates of the extreme ends of all PC axes, e.g. $PC1min

sdev

The standard deviations of the principal components (i.e., the square roots of the eigenvalues of the covariance/correlation matrix, as per prcomp.

rotation

The matrix of variable loadings, as per prcomp.

Details

The function performs a principal components analysis of shape variation and plots two dimensions of tangent space for a set of Procrustes shape variables (default is PC1 vs. PC2). The percent variation along each PC-axis is returned. Additionally (and optionally, warpgrids=T), deformation grids can be requested, which display the shape of specimens at the ends of the range of variability along PC1. If groups are provided, specimens from each group are plotted using distinct colors based on the order in which the groups are found in the dataset, and using R's standard color palette: black, red, green, blue, cyan, magenta, yellow, and gray. NOTE: to change the colors of the groups, simply substitute a vector of the desired colors for each specimen (see example below).

NOTE: previous versions of plotTangentSpace had option 'verbose' to return the PC scores and PC shapes. From version 3.0.2 this is automatic when assigned to an object.

Examples

Run this code

# NOT RUN {
data(plethodon) 
Y.gpa<-gpagen(plethodon$land)    #GPA-alignment

gp <- interaction(plethodon$species, plethodon$site) # group must be a factor
plotTangentSpace(Y.gpa$coords, groups = gp) 

## To save and use output
PCA <- plotTangentSpace(Y.gpa$coords, groups = gp, legend=TRUE) 
summary(PCA)
PCA$pc.shapes
PCA$rotation

##To change colors of groups
col.gp <- rainbow(length(levels(gp))) 
   names(col.gp) <- levels(gp)
col.gp <- col.gp[match(gp, names(col.gp))] # col.gp must NOT be a factor
plotTangentSpace(Y.gpa$coords, groups = col.gp)

## To plot residual shapes from an allometry regression (note: must add mean back in!) 
plotTangentSpace(arrayspecs(resid(lm(two.d.array(Y.gpa$coords)~Y.gpa$Csize))+
         predict(lm(two.d.array(Y.gpa$coords)~1)),12,2))
# }