esetTsne: plot a t-SNE of an eSet object
In esetVis: Visualizations of expressionSet Bioconductor object

Description Usage Arguments Value Author(s) References See Also Examples

esetTsne reduces the dimension of the data contained in the eSet via t-Distributed Stochastic Neighbor Embedding with the Rtsne function and plot the subsequent biplot, possibly with sample annotation contained in the eSet.

esetTsne(eset, psids = 1:nrow(eset), trace = TRUE,
  colorVar = character(), color = if (length(colorVar) == 0) "black"
  else character(), shapeVar = character(), shape = if
  (length(shapeVar) == 0) 15 else numeric(), sizeVar = character(),
  size = if (length(sizeVar) == 0) {     ifelse(typePlot == "interactive"
  && length(packageInteractivity) == 1 && packageInteractivity == "rbokeh",
  5, 2.5) } else {     numeric() }, sizeRange = numeric(),
  alphaVar = character(), alpha = if (length(alphaVar) == 0) 1 else
  numeric(), alphaRange = numeric(), title = "",
  Rtsne.args = list(perplexity = floor((ncol(eset) - 1)/3), theta = 0.5,
  dims = 2, initial_dims = 50), fctTransformDataForInputTsne = NULL,
  symmetryAxes = c("combine", "separate", "none"),
  packageTextLabel = c("ggrepel", "ggplot2"), topSamples = 10,
  topSamplesCex = 2.5, topSamplesVar = character(),
  topSamplesJust = c(0.5, 0.5), topSamplesColor = "black",
  includeLegend = TRUE, includeLineOrigin = TRUE,
  typePlot = c("static", "interactive"),
  packageInteractivity = c("rbokeh", "ggvis"),
  figInteractiveSize = c(600, 400), ggvisAdjustLegend = TRUE,
  interactiveTooltip = TRUE, interactiveTooltipExtraVars = character(),
  returnAnalysis = FALSE, returnEsetPlot = FALSE)

`eset`	expressionSet (or SummarizedExperiment) object with data
`psids`	featureNames of genes to include in the plot, all by default
`trace`	logical, if TRUE (by default), print some messages during tsne is running
`colorVar`	name of variable (in varLabels of the `eset`) used for coloring, empty by default
`color`	character or factor with specified color(s) for the points, replicated if needed. This is used only if `colorVar` is empty. By default: 'black' if `colorVar` is not specified and default `ggplot` palette otherwise
`shapeVar`	name of variable (in varLabels of the `eset`) used for the shape, empty by default
`shape`	character or factor with specified shape(s) (pch) for the points, replicated if needed. This is used only if `shapeVar` is empty. By default: '15' (filled square) if `shapeVar` is not specified and default `ggplot` shape(s) otherwise
`sizeVar`	name of variable (in varLabels of the `eset`) used for the size, empty by default
`size`	character or factor with specified size(s) (cex) for the points, replicated if needed. This is used only if `sizeVar` is empty. By default: '2.5' if `sizeVar` is not specified and default `ggplot` size(s) otherwise
`sizeRange`	size (cex) range used in the plot, possible only if the `sizeVar` is 'numeric' or 'integer'
`alphaVar`	name of variable (in varLabels of the `eset`) used for the transparency, empty by default. This parameter is currently only available for static plot and ggvis (only numeric in this case).
`alpha`	character or factor with specified transparency(s) for the points, replicated if needed. This is used only if `shapeVar` is empty. By default: '1' if `alphaVar` is not specified and default `ggplot` alpha otherwise This parameter is currently only available for static and ggvis.
`alphaRange`	transparency (alpha) range used in the plot, possible only if the `alphaVar` is 'numeric' or 'integer' This parameter is currently only available for static and ggvis plot.
`title`	plot title, ” by default
`Rtsne.args`	arguments for the Rtsne function, by default: perplexite parameter = optimal number of neighbours, theta = speed/accuracy trade-off (increase for less accuracy), set to 0.0 for exact TSNE
`fctTransformDataForInputTsne`	function which transform the data in the eSet object before calling the `Rtsne` function. This should be a function which takes a matrix as input and return a matrix, e.g. the dist function.
`symmetryAxes`	set symmetry for axes, either: 'combine' (by default): both axes are symmetric and with the same limits 'separate': each axis is symmetric and has its own limits 'none': axes by default (plot limits)
`packageTextLabel`	package used to label the outlying genes/samples/gene sets, either `ggrepel` (by default, only used if package `ggrepel` is available), or `ggplot2`
`topSamples`	numeric indicating which percentile (if <1) or number (if >=1) of samples most distant to the origin of the plot to annotate, by default: 10 samples are selected If no samples should be annotated, set this parameter to 0. Currently available for static plot.
`topSamplesCex`	cex for sample annotation (used when `topSamples` > 0)
`topSamplesVar`	variable of the phenoData used to label the samples, by default: empty, the sampleNames are used for labelling (used when `topSample`s > 0)
`topSamplesJust`	text justification for the samples (used when `topSamples` > 0 and if `packageTextLabel` is `ggplot2`), by default: c(0.5, 0.5) so centered
`topSamplesColor`	text color for the samples (used when `topSamples` > 0), black by default
`includeLegend`	logical if TRUE (by default) include a legend, otherwise not
`includeLineOrigin`	if TRUE (by default) include vertical line at x = 0 and horizontal line at y = 0
`typePlot`	type of the plot returned, either 'static' (static) or interactive' (potentially interactive)
`packageInteractivity`	if `typePlot` is 'interactive', package used for interactive plot, either 'rbokeh' (by default) or 'ggvis'
`figInteractiveSize`	vector containing the size of the interactive plot, as [width, height] by default: c(600, 400). This is passed to the `width` and `height` parameters of: for rbokeh plots: the `bokeh::figure` function for ggvis plots: the `ggvis::set_options` function
`ggvisAdjustLegend`	logical, if TRUE (by default) adjust the legends in `ggvis` to avoid overlapping legends when multiple legends
`interactiveTooltip`	logical, if TRUE, add hoover functionality showing sample annotation (variables used in the plot) in the plot
`interactiveTooltipExtraVars`	name of extra variable(s) (in varLabels of the `eset`) to add in rbokehEsetPlot to label the samples, empty by default
`returnAnalysis`	logical, if TRUE (FALSE by default), return also the output of the analysis, and the outlying samples in the topElements element if any, otherwise only the plot object
`returnEsetPlot`	logical, if TRUE return also the esetPlot object

if returnAnalysis is TRUE, return a list:

analysis: output of the spectral map analysis, whose elements can be given to the esetPlotWrapper function
- dataPlotSamples: coordinates of the samples
- esetUsed: expressionSet used in the plot
topElements: list with top outlying elements if any, possibly genes, samples and gene sets
plot: the plot output

otherwise return only the plot

Laure Cougnaud

L.J.P. van der Maaten and G.E. Hinton (2008). Visualizing High-Dimensional Data Using t-SNE. Journal of Machine Learning Research, 2579–2605

the function used internally: Rtsne or http://homepage.tudelft.nl/19j49/t-SNE.html for further explanations about this technique.

library(ALL)
data(ALL)

## complete example (most of the parameters are optional)

# create custom color palette
colorPalette <- c("dodgerblue", colorRampPalette(c("white","dodgerblue2", "darkblue"))(5)[-1], 
	"red", colorRampPalette(c("white", "red3", "darkred"))(5)[-1])

# create tsne
print(esetTsne(eset = ALL, 
	title = "Acute lymphoblastic leukemia dataset \n Tsne complete",
	colorVar = "BT", color = colorPalette,
	shapeVar = "sex", shape = 15:16,
	sizeVar = "age", sizeRange = c(2, 6),
	symmetryAxes = "separate",
	topSamples = 15, topSamplesVar = "cod", topSamplesColor = "black",
	topSamplesJust = c(1, 0), topSamplesCex = 3)
)

Loading required package: Biobase
Loading required package: BiocGenerics
Loading required package: parallel

Attaching package: 'BiocGenerics'

The following objects are masked from 'package:parallel':

    clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
    clusterExport, clusterMap, parApply, parCapply, parLapply,
    parLapplyLB, parRapply, parSapply, parSapplyLB

The following objects are masked from 'package:stats':

    IQR, mad, sd, var, xtabs

The following objects are masked from 'package:base':

    Filter, Find, Map, Position, Reduce, anyDuplicated, append,
    as.data.frame, cbind, colMeans, colSums, colnames, do.call,
    duplicated, eval, evalq, get, grep, grepl, intersect, is.unsorted,
    lapply, lengths, mapply, match, mget, order, paste, pmax, pmax.int,
    pmin, pmin.int, rank, rbind, rowMeans, rowSums, rownames, sapply,
    setdiff, sort, table, tapply, union, unique, unsplit, which,
    which.max, which.min

Welcome to Bioconductor

    Vignettes contain introductory material; view with
    'browseVignettes()'. To cite Bioconductor, see
    'citation("Biobase")', and for packages 'citation("pkgname")'.

Warning message:
Removed 5 rows containing missing values (geom_point).