plottingFunctions: Convert clusterLegend into useful formats

In epurdom/clusterExperiment: Compare Clusterings for Single-Cell Sequencing

Convert clusterLegend into useful formats

Description

Function for converting the information stored in the clusterLegend slot into other useful formats.

Most of these functions are called internally by plotting functions, but are exported in case the user finds them useful.

Usage

makeBlankData(
  data,
  groupsOfFeatures = NULL,
  groupsOfSamples = NULL,
  nBlankFeatures = 1,
  nBlankSamples = 1
)

## S4 method for signature 'ClusterExperiment'
convertClusterLegend(
  object,
  output = c("plotAndLegend", "aheatmapFormat", "matrixNames", "matrixColors"),
  whichClusters = ifelse(output == "plotAndLegend", "primary", "all")
)

showPalette(colPalette = bigPalette, which = NULL, cex = 1)

bigPalette

massivePalette

setBreaks(data, breaks = NA, makeSymmetric = FALSE, returnBreaks = TRUE)

showHeatmapPalettes()

seqPal5

seqPal2

seqPal3

seqPal4

seqPal1

## S4 method for signature 'ClusterExperiment'
plotClusterLegend(
  object,
  whichCluster = "primary",
  clusterNames,
  title,
  add = FALSE,
  location = if (add) "topright" else "center",
  ...
)

Arguments

data	matrix with samples on columns and features on rows.
groupsOfFeatures	list, with each element of the list containing a vector of numeric indices of features (rows).
groupsOfSamples	list, with each element of the list containing a vector of numeric indices of samples (columns).
nBlankFeatures	the number of blank lines to add in the data matrix to separate the groups of feature indices (will govern the amount of white space if data is then fed to heatmap.)
nBlankSamples	the number of blank lines to add in the data matrix to separate the groups of sample indices (will govern the amount of white space if data is then fed to heatmap.)
object	a ClusterExperiment object.
output	character value, indicating desired type of conversion.
whichClusters	argument that can be either numeric or character vector indicating the clusterings to be used. See details of getClusterIndex.
colPalette	a vector of character colors. By default, the palette bigPalette is used
which	numeric. Which colors to plot. Must be a numeric vector with values between 1 and length of colPalette. If missing, all colors plotted.
cex	numeric value giving the cex for the text of the plot.
breaks	either vector of breaks, or number of breaks (integer) or a number between 0 and 1 indicating a quantile, between which evenly spaced breaks should be calculated. If missing or NA, will determine evenly spaced breaks in the range of the data.
makeSymmetric	whether to make the range of the breaks symmetric around zero (only used if not all of the data is non-positive and not all of the data is non-negative)
returnBreaks	logical as to whether to return the vector of breaks. See details.
whichCluster	argument that can be a single numeric or character value indicating the single clustering to be used. Giving values that result in more than one clustering will result in an error. See details of getClusterIndex.
clusterNames	vector of names for the clusters; vector should have names that correspond to the clusterIds in the ClusterExperiment object. If this argument is missing, will use the names in the "name" column of the clusterLegend slot of the object.
title	title for the clusterLegend plot
add	logical. Whether legend should be added to the existing plot.
location	character passed to x argument of legend indicating where to place legend.
...	arguments passed to legend

Format

An object of class character of length 56.

An object of class character of length 484.

An object of class character of length 16.

An object of class character of length 14.

An object of class character of length 11.

An object of class character of length 13.

An object of class character of length 11.

Details

makeBlankData pulls the data corresponding to the row indices in groupsOfFeatures adds lines of NA values into data between these groups. When given to heatmap, will create white space between these groups of features.

convertClusterLegend pulls out information stored in the clusterLegend slot of the object and returns it in useful format.

bigPalette is a long palette of colors (length 58) used by plotClusters and accompanying functions. showPalette creates plot that gives index of each color in a vector of colors. massivePalette is a combination of bigPalette and the non-grey colors of colors() (length 487). massivePalette is mainly useful for when doing plotClusters of a very large number of clusterings, each with many clusters, so that the code doesn't run out of colors. However, many of the colors will be very similar to each other.

showPalette will plot the colPalette colors with their labels and index.

if returnBreaks if FALSE, instead of returning the vector of breaks, the function will just return the second smallest and second largest value of the breaks. This is useful for alternatively just setting values of the data matrix larger than these values to this value if breaks was a percentile. This argument is only used if breaks<1, indicating truncating the breaks for large values of data.

setBreaks gives a set of breaks (of length 52) equally spaced between the boundaries of the data. If breaks is between 0 and 1, then the evenly spaced breaks are between these quantiles of the data.

seqPal1-seqPal4 are palettes for the heatmap. showHeatmapPalettes will show you these palettes.

Value

makeBlankData returns a list with items

• "dataWBlanks" The data with the rows of NAs separating the given indices.

• "rowNamesWBlanks" A vector of characters giving the rownames for the data, including blanks for the NA rows. These are not given as rownames to the returned data because they are not necessarily unique. However, they can be given to the labRow argument of aheatmap or plotHeatmap.

• "colNamesWBlanks" A vector of characters giving the colnames for the data, including blanks for the NA rows. They can be given to the labCol argument of aheatmap or plotHeatmap.

• "featureGroupNamesWBlanks" A vector of characters of the same length as the number of rows of the new data (i.e. with blanks) giving the group name for the data, indicating which group (i.e. which element of groupsOfFeatures list) the feature came from. If groupsOfFeatures has unique names, these names will be used, other wise "Feature Group1", "Feature Group2", etc. The NA rows are given NA values.

• "sampleGroupNamesWBlanks" A vector of characters of the same length as the number of columns of the new data (i.e. with blanks) giving the group name for the data, indicating which group (i.e. which element of groupsOfFeatures list) the feature came from. If groupsOfFeatures has unique names, these names will be used, other wise "SampleGroup1", "Group2", etc. The NA rows are given NA values.

If output="plotAndLegend", "convertClusterLegend" will return a list that provides the necessary information to color samples according to cluster and create a legend for it:

• "colorVector" A vector the same length as the number of samples, assigning a color to each cluster of the primaryCluster of the object.

• "legendNames" A vector the length of the number of clusters of primaryCluster of the object giving the name of the cluster.

• "legendColors" A vector the length of the number of clusters of primaryCluster of the object giving the color of the cluster.

If output="aheatmap" a conversion of the clusterLegend to be in the format requested by aheatmap. The column 'name' is used for the names and the column 'color' for the color of the clusters.

If output="matrixNames" or "matrixColors" a matrix the same dimension of clusterMatrix(object), but with the cluster color or cluster name instead of the clusterIds, respectively.

See Also

plotHeatmap

Examples

data(simData)

x <- makeBlankData(simData[,1:10], groupsOfFeatures=list(c(5, 2, 3), c(20,
34, 25)))
plotHeatmap(x$dataWBlanks,clusterFeatures=FALSE)
showPalette()
showPalette(massivePalette,cex=0.6)
setBreaks(data=simData,breaks=.9)

#show the palette colors
showHeatmapPalettes()

#compare the palettes on heatmap
cl <- clusterSingle(simData, subsample=FALSE,
sequential=FALSE, 
mainClusterArgs=list(clusterFunction="pam", clusterArgs=list(k=8)))

## Not run: 
par(mfrow=c(2,3))
plotHeatmap(cl, colorScale=seqPal1, main="seqPal1")
plotHeatmap(cl, colorScale=seqPal2, main="seqPal2")
plotHeatmap(cl, colorScale=seqPal3, main="seqPal3")
plotHeatmap(cl, colorScale=seqPal4, main="seqPal4")
plotHeatmap(cl, colorScale=seqPal5, main="seqPal5")
par(mfrow=c(1,1))

## End(Not run)

epurdom/clusterExperiment documentation built on April 28, 2024, 8:17 p.m.