plot.RCM: Plot RC(M) ordination result with the help of ggplot2

In CenterForStatistics-UGent/RCM: Fit row-column association models with the negative binomial distribution for the microbiome

Plot RC(M) ordination result with the help of ggplot2

Description

Plot RC(M) ordination result with the help of ggplot2

Usage

## S3 method for class 'RCM'
plot(
  x,
  ...,
  Dim = c(1, 2),
  plotType = c("samples", "species", "variables"),
  samColour = if (is.null(inflVar)) NULL else "Influence",
  taxNum = if (all(plotType == "species") || !is.null(taxRegExp)) {
     ncol(x$X)
 }
    else {
     10
 },
  taxRegExp = NULL,
  varNum = 15,
  arrowSize = 0.25,
  inflDim = 1,
  inflVar = NULL,
  returnCoords = FALSE,
  alpha = TRUE,
  varPlot = NULL,
  colLegend = if (!is.null(inflVar)) paste0("Influence on\n", inflVar,
    "\nparameter \nin dimension", inflDim) else samColour,
  samShape = NULL,
  shapeLegend = samShape,
  samSize = 2,
  scalingFactor = NULL,
  quadDrop = 0.995,
  plotEllipse = TRUE,
  taxaScale = 0.5,
  Palette = if (!all(plotType == "species")) "Set1" else "Paired",
  taxLabels = !all(plotType == "species"),
  taxDots = FALSE,
  taxCol = "blue",
  taxColSingle = "blue",
  nudge_y = 0.08,
  axesFixed = TRUE,
  aspRatio = 1,
  xInd = if (all(plotType == "samples")) c(0, 0) else c(-0.75, 0.75),
  yInd = c(0, 0),
  taxLabSize = 4,
  varLabSize = 3.5,
  alphaRange = c(0.2, 1),
  varExpFactor = 10,
  manExpFactorTaxa = 0.975,
  nPhyl = 10,
  phylOther = c(""),
  legendSize = samSize,
  noLegend = is.null(samColour),
  crossSize = 4,
  contCol = c("orange", "darkgreen"),
  legendLabSize = 15,
  legendTitleSize = 16,
  axisLabSize = 14,
  axisTitleSize = 16,
  plotPsi = "psi",
  breakChar = "\n"
)

Arguments

x	an RCM object
...	further arguments, passed on to aes in the the ggplot() function
Dim	An integer vector of length two, which dimensions to plot
plotType	a character string: which components to plot. Can be any combination of 'samples','species' and 'variables'
samColour	a character string, the variable to use for the colour of the sample dots. Can also be a richness measure, or "influence". Alternatively, a vector equal to the number of samples in the RCM object can be supplied. See details.
taxNum	an integer, the number of taxa to be plotted
taxRegExp	a character vector indicating which taxa to plot. Any taxa matcing this regular expression will be plotted
varNum	an integehr, number of variable arrows to draw
arrowSize	a scalar, the size of the arrows
inflDim	an integer, the dimension for which the influence should be calculated
inflVar	the variable on which the influence should be plotted. See details.
returnCoords	a boolean, should final coordinates be returned?
alpha	a boolean, should small arrows be made transparent?
varPlot	the names of the variable arrows to plot. Overrides the varNum argument
colLegend	a character string, the legend text for the sample colour. Defaults to the name of the colour variable
samShape	a character string, the variable to use for the shape of the sample dots
shapeLegend	a character string, the text to use for the shapeLegend. Defaults to the name of the shape variable
samSize	a scalar, the size of the sample dots
scalingFactor	a scalar, a user supplied scaling factor for the taxon arrows. If not supplied it will be calculated to make sample and taxon plots on the same scale
quadDrop	a number between 0 and 1. At this fraction of the peak height are the ellipses of the quadratic response functions drawn
plotEllipse	a boolean, whether to add the ellipses
taxaScale	a scalar, by which to scale the rectangles of the quadratic taxon plot
Palette	the colour palette
taxLabels	a boolean, should taxon labels be plotted?
taxDots	a boolean, should taxa be plotted as dots?
taxCol	the taxon colour
taxColSingle	the taxon colour if there is only one
nudge_y	a scalar, the offet for the taxon labels
axesFixed	A boolean, should the aspect ratio of the plot (the scale between the x and y-axis) be fixed. It is highly recommended to keep this argument at TRUE for honest representation of the ordination. If set to FALSE, the plotting space will be optimally used but the plot may be deformed in the process.
aspRatio	The aspect ratio of the plot when 'axesfixed' is TRUE (otherwise this argument is ignored), passde on to ggplot2::coord_fixed(). It is highly recommended to keep this argument at 1 for honest representation of the ordination.
xInd	a scalar or a vector of length 2, specifying the indentation left and right of the plot to allow for the labels to be printed entirely. Defaults to 0.75 at every side
yInd	a scalar or a vector of length 2, specifying the indentation top and bottom of the plot to allow for the labels to be printed entirely. Defaults to 0 at every side
taxLabSize	the size of taxon labels
varLabSize	the size of the variable label
alphaRange	The range of transparency
varExpFactor	a scalar, the factor by which to expand the variable coordinates
manExpFactorTaxa	a manual expansion factor for the taxa. Setting it to a high value allows you to plot the taxa around the samples
nPhyl	an integer, number of phylogenetic levels to show
phylOther	a character vector of phylogenetic levels to be included in the 'other' group
legendSize	a size for the coloured dots in the legend
noLegend	a boolean indicating you do not want a legend
crossSize	the size of the central cross
contCol	a character vector of length two, giving the low and high values of the continuous colour scale
legendLabSize	size of the legend labels
legendTitleSize	size of the legend title
axisLabSize	size of the axis labels
axisTitleSize	size of the axis title
plotPsi	a character vector, describing what to plot on the axis. Can be either 'psi', 'none' or 'loglik'. The latter plots the log-likelihood explained
breakChar	a character string indicating how the taxon names should be broken

Details

This function relies on the ggplot2 machinery to produce the plots, and the result can be modified accordingly. Monoplots, biplots and for constrained analysis even triplots can be produced, depending on the 'plotType' argument.

When one of either 'Observed', 'Chao1', 'ACE', 'Shannon', 'Simpson', 'InvSimpson' or 'Fisher' are supplied to the 'samColour' argument, the according richness measure (as calculated by phyloseq::estimate_richness) is mapped to the sample colour. When "influence" is supplied, the influence on the variable supplied is plotted. This 'inflVar' variable should be either "psi", or a variable name.

Value

plots a ggplot2-object to output

Note

Supplying only few categorical variables as constraining variables may cause the samples to be plotted on top of each other, since the number of unique sample scores is limited. The plot is still valid, but consider adding more sample variables to spread out the samples

See Also

RCM,addOrthProjection, extractCoord,plotRespFun

Examples

data(Zeller)
require(phyloseq)
tmpPhy = prune_taxa(taxa_names(Zeller)[1:100],
prune_samples(sample_names(Zeller)[1:50], Zeller))
# Subset for a quick fit
zellerRCM = RCM(tmpPhy)
plot(zellerRCM)

CenterForStatistics-UGent/RCM documentation built on April 24, 2023, 8:26 p.m.