coverage_heatmap: coverage_heatmap
In skummerf/GenomicWidgets: interactive visualization of genomics data

Description Usage Arguments Details Value Author(s) Examples

Makes an interactive coverage heatmap.

coverage_heatmap(data, ...)

add_coverage_heatmap(p, data, ...)

## S4 method for signature 'SummarizedExperiment'
coverage_heatmap(data,
  assay = assayNames(data), ...)

## S4 method for signature 'IheatmapHorizontal,SummarizedExperiment'
add_coverage_heatmap(p,
  data, assay = assayNames(data), ...)

## S4 method for signature 'ScoreMatrix'
coverage_heatmap(data, start, end, x = seq(start, end,
  length.out = ncol(data)), y = default_y(data), ...)

## S4 method for signature 'IheatmapHorizontal,ScoreMatrixList'
add_coverage_heatmap(p, data,
  start, end, x = seq(start, end, length.out = ncol(data[[1]])), ...)

## S4 method for signature 'matrix'
coverage_heatmap(data, x = default_x(data),
  y = default_y(data), row_order = c("signal", "hclust", "kmeans", "groups",
  "none"), k = NULL, groups = NULL, clust_dist = stats::dist,
  signal = log10rowMeans(data), plot_signal = TRUE, name = "Coverage",
  signal_name = "Avg. (log10)", summary = TRUE,
  scale_method = c("localRms", "localMean", "localNonZeroMean",
  "PercentileMax", "scalar", "none"), pct = 0.95, scale_factor = 1,
  show_xlabels = TRUE, start = x[1], end = default_end(x),
  col_title = "Position", layout = list(font = list(size = 10)), ...)

## S4 method for signature 'IheatmapHorizontal,matrix'
add_coverage_heatmap(p, data,
  x = default_x(data), signal = log10rowMeans(data), plot_signal = TRUE,
  name = "Coverage", signal_name = "Avg. (log10)", summary = TRUE,
  scale_method = c("localRms", "localMean", "localNonZeroMean",
  "PercentileMax", "scalar", "none"), pct = 0.95, scale_factor = 1,
  show_xlabels = TRUE, start = x[1], end = default_end(x),
  col_title = "Position", ...)

## S4 method for signature 'list'
coverage_heatmap(data, x = default_x(data[[1]]),
  y = default_y(data[[1]]), row_order = c("signal", "hclust", "kmeans",
  "groups", "none"), k = NULL, groups = NULL, clust_dist = stats::dist,
  cluster_by = c("first", "all"), signal = lapply(data, log10rowMeans),
  plot_signal = TRUE, name = "Coverage", signal_name = "Avg. (log10)",
  summary = TRUE, scale_method = c("localRms", "localMean",
  "localNonZeroMean", "PercentileMax", "scalar", "none"), pct = 0.95,
  scale_factor = 1, show_xlabels = TRUE, start = x[1],
  end = default_end(x), col_title = "Position", layout = list(font =
  list(size = 10)), ...)

## S4 method for signature 'IheatmapHorizontal,list'
add_coverage_heatmap(p, data,
  x = default_x(data[[1]]), signal = lapply(data, log10rowMeans),
  plot_signal = TRUE, name = "Coverage", signal_name = "Avg. (log10)",
  summary = TRUE, scale_method = c("localRms", "localMean",
  "localNonZeroMean", "PercentileMax", "scalar", "none"), pct = 0.95,
  scale_factor = 1, show_xlabels = TRUE, start = x[1],
  end = default_end(x), col_title = "Position", ...)

`data`	single coverage matrix or list of coverage matrices
`...`	additional arguments
`p`	IHeatmap object
`assay`	name(s) of assay to plot, if data is SummarizedExperiment
`start`	label for start of x range
`end`	label for end of x range
`x`	x axis labels
`y`	y axis labels
`row_order`	row order method
`k`	k to use for kmeans clustering or cutting heirarchical clustering
`groups`	pre-determined groups for rows
`clust_dist`	distance function to use for clustering
`signal`	signal along row
`plot_signal`	add an annotation heatmap showing the average signal? default is TRUE
`name`	name of colorbar
`signal_name`	name for signal colorbar
`summary`	make summary plot, boolean, default is TRUE
`scale_method`	how to scale matrix before displaying in heatmap, see Details section
`pct`	percentile to use if scale_method is "PercentileMax"
`scale_factor`	scale_factor to use if scale_method is "scalar"
`show_xlabels`	show xlabels? default is TRUE
`col_title`	x axis label
`layout`	list of layout attributes
`cluster_by`	use "first" or "all" matrices for clustering if given multiple matrices

scale_method choices are "localRms", "localMean", "localNonZeroMean", "PercentileMax", "scalar", and "none". localRMS will divide each row by the root mean squared values of that row. localMean will divide each row by the mean of that row. localNonZeroMean will divide each row by nonzero values in that row. PercentileMax will divide values based on percentile (given by pct argument) of the entire matrix. scalar will divide entire matrix by a scalar, given by scalar argument. This scalar could for example be a measure of the sequencing depth.

iheatmap object

Alicia Schep

library(GenomicRanges)
## First we'll read in some sample data
genomation_dir <- system.file("extdata", package = "genomationData")
samp.file <- file.path(genomation_dir,'SamplesInfo.txt')
samp.info <- read.table(samp.file, header=TRUE, sep='\t', 
                        stringsAsFactors = FALSE)
samp.info$fileName <- file.path(genomation_dir, samp.info$fileName)
ctcf.peaks = genomation::readBroadPeak(system.file("extdata",
                          "wgEncodeBroadHistoneH1hescCtcfStdPk.broadPeak.gz",
                           package = "genomationData"))
ctcf.peaks = ctcf.peaks[seqnames(ctcf.peaks) == "chr21"]
ctcf.peaks = ctcf.peaks[order(-ctcf.peaks$signalValue)]
ctcf.peaks = resize(ctcf.peaks, width = 501, fix = "center")

## Make coverage matrix
ctcf_mats <- make_coverage_matrix(samp.info$fileName[1:5], 
                                  ctcf.peaks, 
                                  input_names = samp.info$sampleName[1:5],
                                  up = 250, 
                                  down = 250, 
                                  binsize = 25)
                                  
## Plot coverage for Ctcf and Znf143
if (interactive()){
  coverage_heatmap(ctcf_mats, "Ctcf") %>% 
    add_coverage_heatmap(ctcf_mats, "Znf143")
}