plotAlongChrom: Plot signals and segmentation for a region of a chromosome

In tilingArray: Transcript mapping with high-density oligonucleotide tiling arrays

Description

Plot signals and segmentation for a region of a chromosome

Usage

plotAlongChrom(segObj, y, probeAnno, gff,
    isDirectHybe=FALSE, 
    what = c("dots"), ## "heatmap"
    chr, coord, highlight,
    colors, doLegend=FALSE,
    featureExclude=c("chromosome", "nucleotide_match", "insertion"),
    featureColorScheme=1, extras, 
    rowNamesHeatmap, rowNamesExtras, ylab, ylabExtras, main, 
    colHeatmap=colorRamp(brewer.pal(9, "YlGnBu")), 
    colExtras=colorRamp(brewer.pal(9, "Reds")), 
    sepPlots=FALSE, reOrder=TRUE, ...)

Arguments

segObj	Either an environment or an object of S4 class segmentation. See Details.
y	a numeric vector or matrix containing the signal to be plotted. See Details.
probeAnno	environment with probe annotations. See Details, and package davidTiling for an example.
gff	data frame with genome annotation from the GFF file.
isDirectHybe	logical scalar: if TRUE, the mapping of probes to genomic strands is reversed with respect to the default. This is appropriate for data from a direct RNA hybridization that used no reverse transcription.
what	character scalar indicating which signal visualization to plot. Can be either dots to plot each probe intensity with a point, or heatmap to produce a colorscale representation of the intesities.
chr	integer of length 1 indicating the chromosome number to plot.
coord	integer vector of length 2 containing the start and end coordinates (in bases) for the plot.
highlight	(optional) list with two elements: a single numeric value coord and a character strand. If present, this position is marked by a vertical red bar on the coordinate axis. The color can be changed using the colors argument below.
colors	(optional) named character vector. If missing, a default color scheme is used: c("+"="#00441b", "-"="#081d58", "duplicated"="grey", "cp"="#101010", "highlight"="red", "threshold"="grey"), where the first three elements refer to the colors of data points and the last three to the colors of lines in the plot.
doLegend	logical: should the plot contain a legend?
featureExclude	character vector of names of feature types (in gff) that should not be plotted. Default is "chromosome", "nucleotide_match" and "insertion". Additional possibilities include: "ARS", "repeat\_region", "repeat\_family" and "nc\_primary\_transcript".
featureColorScheme	numeric scalar, used to select a color scheme for the boxes representing genomic features such as coding sequences, ncRNAs etc. Currently the only value supported is 1 (see plotAlongChromLegend or plotFeatures for further information).
extras	a matrix containing additional values to be plotted along the chromosome in a separate panel (such as p-values). This option is only available when y is specified. These values should be on the scale [0,1].
rowNamesHeatmap	character vector of row names for the main heatmap.
rowNamesExtras	character vector of row names for the extra heatmap.
ylab	character label for y-axis of main plot.
ylabExtras	character label for y-axis on extras panel (if specified).
main	character: plot title.
colHeatmap	function describing color scheme for the main heatmap plot (defaults to YlGnBu from RColorBrewer package).
colExtras	function describing color scheme for the extra heatmap plot (if specified) (defaults to Reds from RColorBrewer package).
sepPlots	logical scalar. If TRUE, each column of intensities in segObj or y is plotted separately (maximum of 3) in the same figure. When FALSE, the average is plotted. This argument is only used when what is set to dots.
reOrder	logical scalar (only used when sepPlots is TRUE). If TRUE, the first column of intensities is printed at the bottom of each plot, and the subsequent columns are plotted above. If FALSE, the first appears at the top, and the subsequent columns are plotted below.
...	further arguments that can be passed to the functions that implement the what option above (see plotSegmentationDots and plotSegmentationHeatmap) or gff plotting (see plotFeatures and plotAlongChromLegend).

Details

Intensities: There are two alternative, mutually exclusive ways of providing the intensities that are to be plotted to this function.

1. Via the parameters y and probeAnno. In this case, y is a matrix of intensities, whose rows correspond to probes on the array, and its columns to different conditions, time points, etc. It is also acceptable that y is provided as a vector, in which case it is converted to an nrow(y) x 1 matrix. probeAnno is an environment whose elements correspond to target sequences (e.g. chromosome strands) and that contain integer vectors of length nrow(y) with information about the probes: start and end positions of their alignment to the target sequence, their row indices in y, the type of alignment (is it perfect? is is unique?). For example, the start positions and indices of probes for the + strand of chromosome 1 would be described by environment elements "1.+.start" and "1.+.index".

2. Via the parameter segObj.

segObj: This can be either an object of S4 class segmentation or an environment that by convention contains a certain set of objects. Future work on this package will focus on the S4 class segmentation. The environment option is provided for backward compatibility.

Explanation of the environment: the intended workflow is as follows: Use the script segment.R (in the inst/scripts directory of this package) to generate segmentations. This can be run in parallel on several processors, separately for each chromosome and strand. The results of this are stored in files of the name 1.+.rda, 1.-.rda, 2.+.rda, and so forth, typically within a dedicated directory. Then use the script readSegments.R to collect the R objects in these .rda files into the environment. It contains three types of data:

• microarray intensities in along-chromosome order.

• the segmentation objects (output of findSegments).

• a dataframe named segScore with segment scores; it can be missing iff nrBasesPerSeg is present,

• a numeric scalar names theThreshold, which is used to draw a horizontal "threshold" line in the plot.

... and the different signal visualization methods (what option): If what=="dots", the argument showConfidenceIntervals can be a logical scalar to choose whether vertical dashed lines are drawn for the confidence interval. In any case, these are only drawn if they are present in the segmentation object in segObj.

Author(s)

Wolfgang Huber <[e-mail]>

Examples

  ## 1. see viewSegmentation.R script in the inst/scripts directory
  ## 2. (newer): segmentation.Rnw
  ## 3. (newer): see the plotALongChrom vignette
  data(segnf)
  data(gffSub)
  nmLabel = colnames(segnf$"1.+"@y)
  plotAlongChrom(segnf,chr=1,coord=c(35000,50000),
  			gff=gffSub,rowNamesHeatmap=nmLabel) ##the dots
  plotAlongChrom(segnf,chr=1,coord=c(35000,50000),what="heatmap", 
  			gff=gffSub,rowNamesHeatmap=nmLabel) ##the heatmap
  			
  plotAlongChrom(segnf,chr=1,coord=c(35000,50000),gff=gffSub,
  			showConfidenceIntervals=FALSE) ##do not show the segment confidence interval	
   

tilingArray documentation built on Nov. 8, 2020, 10:59 p.m.