ggplot Method
This method uses the ggplot2 package to draw highly
customizable ToxPi profiles with multiple different aesthetics and
information options. The most notable aesthetic addition is the
inclusion of information regarding the amount of missing data each slice
has, which is represented by the a small circle of interior slices as is
shown in profiles generated by the ToxPi GUI. Each interior slice is
colored on a grey scale, with black representing all missing data and
white representing no missing data. Each new customization option, along
with some other notable differences, are described below.
Customization Options (Arguments)
fills : Vector containing slice colors. Defaults to a color
scheme matching the ToxPi GUI.
showMissing : Boolean to show missing data information.
Defaults to TRUE.
showCenter : Boolean to show inner circle. Defaults to TRUE.
FALSE overrides showMissing.
showScore : Boolean to show overall profile scores. Defaults
to TRUE.
ncol : Integer specifying number of columns in plot. Default
is an automatic calculation by ggplot.
bgColor : Text specifying background color for profiles as
either hex codes or R recognized colors. Defaults to “gray80”.
borderColor : Text specifying color of max radius ring as
either hex codes or R recognized colors. Defaults to “white”.
sliceBorderColor : Text specifying color of slice borders as
either hex codes or R recognized colors. Defaults to “white”.
sliceValueColor : Text specifying color of slice scores as
either hex codes or R recognized colors. Defaults to NULL.
sliceLineColor : Text specifying color of slice guidelines
as either hex codes or R recognized colors. Defaults to NULL.
Usage
Depending on export method, plot resolution may end up poor. SVG and
PDF images tend to produce better results than PNG. Usage and example
plot output are shown below.
library("ggplot2")
# Default plot
plot(f.results, package = "gg") # Using ggplot package
Sometimes the inner circle of the ToxPi figure can show plotting
artifacts. A possible workaround for this is to create a wrapper
function around ggplot2:::coord_munch as shown below.
## Optional code to make smoother lines
# Plot before updating coord_munch
plot(f.results["chem01"], package = "gg")
# Save the original version of coord_munch
coord_munch_orig <- ggplot2:::coord_munch
# Make a wrapper function that has a different default for segment_length
if (length(formals(coord_munch_orig)) == 5) {
coord_munch_new <- function(coord, data, range, segment_length = 1/1000,
is_closed = FALSE) {
coord_munch_orig(coord, data, range, segment_length, is_closed)
}
} else {
coord_munch_new <- function(coord, data, range, segment_length = 1/1000) {
coord_munch_orig(coord, data, range, segment_length)
}
}
# The environment may need to be set
#environment(coord_munch_new) <- environment(coord_munch_orig)
# Replace ggplot2:::coord_munch with coord_munch_new
assignInNamespace("coord_munch", coord_munch_new, ns = "ggplot2")
# Plot after updating coord_munch
plot(f.results["chem01"], package = "gg")
# Revert to original coord_munch ater plotting if desired
#assignInNamespace("coord_munch", coord_munch_orig, ns = "ggplot2")
Customization examples
Coloration Options
# Changing the slice colors
colors <- c("orange", "green", "magenta", "lightblue")
plot(f.results["chem02"], package = "gg", fills = colors)
plot(f.results["chem02"], package = "gg", fills = NULL)
# Changing the background color
plot(f.results["chem02"], package = "gg", bgColor = "lightskyblue")
plot(f.results["chem02"], package = "gg", bgColor = NULL)
# Changing the max radius ring color
plot(f.results["chem02"], package = "gg", borderColor = "black")
plot(f.results["chem02"], package = "gg", borderColor = NULL)
# Changing the slice border color
plot(f.results["chem02"], package = "gg", sliceBorderColor = "magenta")
plot(f.results["chem02"], package = "gg", sliceBorderColor = NULL)
# Adding slice guidelines
plot(f.results["chem02"], package = "gg", sliceLineColor = "red")
plot(f.results["chem02"], package = "gg", sliceLineColor = NULL)
# Adding visible slice scores
plot(f.results["chem02"], package = "gg", sliceValueColor = "brown")
plot(f.results["chem02"], package = "gg", sliceValueColor = NULL)
Binary Options
# Hiding inner circle
plot(f.results["chem02"], package = "gg", showCenter = TRUE)
plot(f.results["chem02"], package = "gg", showCenter = FALSE)
# Hiding missing data information (pure white inner circle)
plot(f.results["chem02"], package = "gg", showMissing = TRUE)
plot(f.results["chem02"], package = "gg", showMissing = FALSE)
# Hiding the overall profile scores
plot(f.results["chem02"], package = "gg", showScore = TRUE)
plot(f.results["chem02"], package = "gg", showScore = FALSE)
pieGrob Method
This method is the default and uses the grid package with grob
objects to draw ToxPi profiles that can be individually customized after
plotting. This method does not allow for the wide array of aesthetics as
ggplot does, but instead it allows the user to highlight specific
profiles or slices that have significant importance in their results.
Methods for highlighting information after plotting is shown below.
library(grid) # Load library
plot(f.results) # ToxPi visuals
# grid.ls() #List grid info
# Highlight one figure using its label
grid.edit("pie-1", fills = c("red", "blue", "black", "brown"))
# Or just one slice in a figure
grid.edit("pie-10::slice1", gp = gpar(fill = "#FC0FC0"))
ggplot vs grid Comparison
#Single sample
plot(f.results["chem02"])
plot(f.results["chem02"], package = "gg")
# Subset plots
plot(f.results[order(txpRanks(f.results))[1:4]]) #Profiles ranked 1-4
plot(f.results[order(txpRanks(f.results))[1:4]], package = "gg") #Profiles ranked 1-4
## Long sample names for cramped plots
#change the first sample name in f.results
txpIDs(f.results)[1] <- "I am a long sample name"
plot(f.results) #grid plot for all samples
plot(f.results, package = "gg") #ggplot for all samples
txpIDs(f.results)[1] <- "chem01" # Change the sample name back
## Long slice names for cramped plots
#change first slice name in model slot
names(f.results@txpModel)[1] <- "long slice name"
#change first slice name in scores slot
colnames(f.results@txpSliceScores)[1] <- "long slice name"
#change first slice name in missing data slot
names(f.results@txpMissing)[1] <- "long slice name"
#plot results using grid
plot(f.results)
#plot results using ggplot
plot(f.results, package = "gg") + theme(
legend.position = "bottom",
legend.title = element_text(size = 10),
legend.text = element_text(size = 6)
)
#change slice name back
names(f.results@txpModel)[1] <- "s1"
colnames(f.results@txpSliceScores)[1] <- "s1"
names(f.results@txpMissing)[1] <- "s1"
Statistic Plots
# Rank plot
plot(f.results, y = txpRanks(f.results), labels = 1:10)
# Hierarchical Clustering
f.hc <- hclust(dist(txpSliceScores(f.results)))
plot(f.hc, hang = -1, labels = txpIDs(f.results), xlab = '', sub = '')
