redartR is a builder for PCoA plots from the dartR package. The
original dartR::gl.pcoa.plot function builds the ggplot object
internally and returns it to the user, but it can then sometimes be
difficult to modify it to suit your needs.
This gl.pcoa.plot.builder function takes a slightly different
approach, returning an object of class ggbuilder, which contains the
data and code required to produce the ggplot object. You can modify
that code and/or data inside the ggbuilder object. Printing or
plotting the ggbuilder object then evaluates that code and produces a
ggplot object; printing or plotting that object will draw the plot
to the current graphics device, as per any other ggplot object.
## load the dartR packge
library(dartR)
## and the functions from this repo
source("redartR.R")
## example plot
gl <- testset.gl
levels(pop(gl)) <- c(rep("Coast", 5), rep("Cooper", 3),rep("Coast", 5),
rep("MDB", 8), rep("Coast", 7), "Em.subglobosa",
"Em.victoriae")
pcoa <- gl.pcoa(gl, nfactors = 5)
#> Performing a PCoA, individuals as entities, SNP loci as attributes
#> Ordination yielded 14 informative dimensions from 249 original dimensions
#> PCoA Axis 1 explains 23.3 % of the total variance
#> PCoA Axis 1 and 2 combined explain 42.8 % of the total variance
#> PCoA Axis 1-3 combined explain 54.4 % of the total varianceThe existing gl.pcoa.plot function:
gl.pcoa.plot(pcoa, gl, ellipse = TRUE, p = 0.95, labels = "pop", hadjust = 1.5, vadjust = 1)
#> Plotting populations
Construct the ggbuilder object:
xb <- gl.pcoa.plot.builder(pcoa, gl, ellipse = TRUE, p = 0.95, labels = "pop", hadjust = 1.5, vadjust = 1)
class(xb)
#> [1] "ggbuilder"This object contains a bunch of elements of class ggplotter, along
with a plot_sequence element:
str(xb, max.level = 2)
#> List of 9
#> $ init :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ points :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ labels :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ theme :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ axis_labels :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ hline :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ vline :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ ellipse :List of 1
#> ..$ :List of 2
#> .. ..- attr(*, "class")= chr "ggplotter"
#> $ plot_sequence: chr [1:8] "init" "points" "labels" "theme" ...
#> - attr(*, "class")= chr "ggbuilder"Each element of class ggplotter contains the data and code required to
produce one part of the plot, and the plot_sequence elements gives the
order in which those elements should be executed.
Look at one of the ggplotter objects:
str(xb$points)
#> List of 1
#> $ :List of 2
#> ..$ plotfun : chr "ggplot2::geom_point"
#> ..$ plotargs:List of 1
#> .. ..$ size: num 2
#> ..- attr(*, "class")= chr "ggplotter"It contains a plotfun element (the function that will be called to
plot this element) and plotargs (the parameters that will be passed to
that function).
Plotting the xb object produces a ggplot object:
xp <- plot(xb)
class(xp)
#> [1] "gg" "ggplot"And printing or plotting the xp object produces the actual plot:
xp
The whole point of this approach is that the plot can be manipulated by
changing the ggplotter elements. We can modify the code or data in
these elements before calling plot(xb) to generate the actual
ggplot. For example, let’s replace the ggplotter element that
controls the points with one that has a shape aesthetic:
xb$points <- as_plotter(plotfun = "ggplot2::geom_point",
plotargs = list(mapping = aes_string(shape = "pop"), size = 2))And now draw the modified plot:
xb
This is largely just a proof of concept. Only labels of type "pop",
"ind", "interactive", or "plotly" are supported at the moment.
This idea draws on a similar approach used in the SOmap package.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent