ScoreMatrixBin-methods: Get bin score for bins on each window
In genomation: Summary, annotation and visualization of genomic data

Description Usage Arguments Value See Also Examples

The function first bins each window to equal number of bins, and calculates the a summary matrix for scores of each bin (currently, mean, max and min supported) A scoreMatrix object can be used to draw average profiles or heatmap of read coverage or wig track-like data. windows can be a predefined region such as CpG islands, gene bodies, transcripts or CDS (coding sequences) that are not necessarily equi-width. Each window will be chopped to equal number of bins based on bin.num option.

ScoreMatrixBin(target, windows, bin.num = 10, bin.op = "mean",
  strand.aware = FALSE, weight.col = NULL, is.noCovNA = FALSE,
  type = "auto", rpm = FALSE, unique = FALSE, extend = 0,
  param = NULL, bam.paired.end = FALSE, library.size = NULL)

\S4method{ScoreMatrixBin}{RleList,GRanges}(target, windows, bin.num, bin.op,
                                                   strand.aware)

\S4method{ScoreMatrixBin}{GRanges,GRanges}(target,windows,
                                                   bin.num,bin.op,
                                                   strand.aware,weight.col,
                                                   is.noCovNA)

\S4method{ScoreMatrixBin}{character,GRanges}(target, windows, bin.num=10,
                                                     bin.op='mean',strand.aware, 
                                                     is.noCovNA=FALSE, type='auto',
                                                     rpm, unique, extend, param,
                                                     bam.paired.end=FALSE, 
                                                     library.size=NULL)

\S4method{ScoreMatrixBin}{RleList,GRangesList}(target,windows,
                                                        bin.num, bin.op,
                                                        strand.aware)

\S4method{ScoreMatrixBin}{GRanges,GRangesList}(target,windows,
                                                   bin.num,bin.op,
                                                   strand.aware,weight.col,
                                                   is.noCovNA)

\S4method{ScoreMatrixBin}{character,GRangesList}(target, windows, bin.num=10,
                                                     bin.op='mean',strand.aware, 
                                                     weight.col=NULL,
                                                     is.noCovNA=FALSE, type='auto',
                                                     rpm, unique, extend, param,
                                                     bam.paired.end=FALSE, 
                                                     library.size=NULL)

`target`	`RleList`, `GRanges`, a BAM file or a bigWig file object to be overlapped with ranges in `windows`
`windows`	`GRanges` or `GRangesList` object that contains the windows of interest. It could be promoters, CpG islands, exons, introns as GRanges object or GrangesList object representing exons of each transcript. Exons must be ordered by ascending rank by their position in transcript. The sizes of windows does NOT have to be equal.
`bin.num`	single `integer` value denoting how many bins there should be for each window
`bin.op`	bin operation that is either one of the following strings: "max","min","mean","median","sum". The operation is applied on the values in the bin. Defaults to "mean"
`strand.aware`	If TRUE (default: FALSE), the strands of the windows will be taken into account in the resulting `scoreMatrix`. If the strand of a window is -, the values of the bins for that window will be reversed
`weight.col`	if the object is `GRanges` object a numeric column in meta data part can be used as weights. This is particularly useful when genomic regions have scores other than their coverage values, such as percent methylation, conservation scores, GC content, etc.
`is.noCovNA`	(Default:FALSE) if TRUE,and if 'target' is a GRanges object with 'weight.col' provided, the bases that are uncovered will be preserved as NA in the returned object. This useful for situations where you can not have coverage all over the genome, such as CpG methylation values.
`type`	(Default:"auto") if target is a character vector of file paths, then type designates the type of the corresponding files (bam or bigWig)
`rpm`	boolean telling whether to normalize the coverage to per milion reads. FALSE by default. See `library.size`.
`unique`	boolean which tells the function to remove duplicated reads based on chr, start, end and strand
`extend`	numeric which tells the function to extend the reads to width=extend
`param`	ScanBamParam object
`bam.paired.end`	boolean indicating whether given BAM file contains paired-end reads (default:FALSE). Paired-reads will be treated as fragments.
`library.size`	numeric indicating total number of mapped reads in a BAM file (`rpm` has to be set to TRUE). If is not given (default: NULL) then library size is calculated using the Rsamtools idxstatsBam function: sum(idxstatsBam(target)$mapped).

returns a scoreMatrix object

ScoreMatrix

data(cage)
data(cpgi)
data(promoters)
myMat=ScoreMatrixBin(target=cage,
                      windows=cpgi,bin.num=10,bin.op="mean",weight.col="tpm")

plot(colMeans(myMat,na.rm=TRUE),type="l")

  
myMat2=ScoreMatrixBin(target=cage,
                       windows=promoters,bin.num=10,bin.op="mean",
                       weight.col="tpm",strand.aware=TRUE)

plot(colMeans(myMat2,na.rm=TRUE),type="l")


# Compute transcript coverage of a set of exons.
library(GenomicRanges)
bed.file = system.file("extdata/chr21.refseq.hg19.bed", 
                       package = "genomation")
gene.parts = readTranscriptFeatures(bed.file)
transcripts = split(gene.parts$exons, gene.parts$exons$name)
transcripts = transcripts[]
myMat3 = ScoreMatrixBin(target=cage, windows=transcripts[1:250], 
                    bin.num=10)
myMat3