RoarDatasetMultipleAPA-constructor: Creates a 'RoarDatasetMultipleAPA' object
In vodkatad/roar: Identify differential APA usage from RNA-seq alignments

Description Usage Arguments Value See Also Examples

This function creates an RoarDatasetMultipleAPA object from two lists of of GAlignments and a GRanges containing a suitable annotation of alternative APA sites and gene exon structure. A MultipleAPA analysis computes several roar values and p-values for each gene: one for every possible combination of APA-canonical end of a gene (i.e. the end of its last exon). This is more efficient than performing several different “standard” roar analyses choosing the PRE and POST portions corresponding to different APAs because reads overlaps are computed only once.

1 2	RoarDatasetMultipleAPA(treatmentBamsGenomicAlignments, controlBamsGenomicAlignments, gtfGRanges)

`treatmentBamsGenomicAlignments`	A list of `GAlignments` representing alignment of samples for the treatment condition (by convention it is considered the “treated” condition: this simply means that the package will compute roar values (ratios of the m/M) using this condition as the numerator) to be considered.
`controlBamsGenomicAlignments`	A list of `GAlignments` representing alignment of samples for the control condition to be considered.
`gtfGRanges`	A `GRanges` containing a suitable annotation of alternative APA sites and gene exonic structure. Minimal requirements are: metadata columns called "gene", "apa" and "type." APA should be single bases falling over one of the given genes and need to have the metadata column "type" equal to "apa" and the "apa" column composed of unambiguous id and the corresponding gene id pasted together with an underscore. The "gene" metadata columns for these entries should not be initialized. All the studied gene exons need to be reported, in this case the metadata column "gene" should contain the gene id (the same one reported for each gene APAs) while "type" should be set to "gene" and "apa" to NA. All apa entries assigned to a gene should have coordinates that falls inside it and every gene that appears should contain at least one APA.

A RoarDatasetMultipleAPA object ready to be analyzed via the other methods.

RoarDatasetMultipleAPAFromFiles

   
      library(GenomicAlignments)
      gene <- c("A", "B", NA, NA)
      type <- c("gene","gene","apa", "apa")
      apa <- c(NA, NA, "apa1_A", "apa2_B")
      features <- GRanges(
         seqnames = Rle(c("chr1", "chr2", "chr1", "chr2")),
         strand = strand(rep("+", length(gene))),
         ranges = IRanges(
            start=c(1000, 2000, 1300, 2050),
            width=c(500, 900, 1, 1)),
         DataFrame(gene, apa, type)
      )
      rd1 <- GAlignments("a", seqnames = Rle("chr1"), pos = as.integer(1000), cigar = "300M", strand = strand("+"))
      rds <- RoarDatasetMultipleAPA(list(c(rd1,rd1)), list(c(rd1,rd1)), features)