computeRoars-method: Computes m/M and roar values

In vodkatad/roar: Identify differential APA usage from RNA-seq alignments

Description

This is the second step in the Roar analyses: it computes the ratio of prevalence of the short and long isoforms for every gene in the treatment and control condition (m/M) and their ratio, roar, that indicates if there is a relative shortening-lengthening in a condition over the other one. A roar > 1 for a given gene means that in the treatment condition that gene has an higher ratio of short vs long isoforms with respect to the control condition (and the opposite for roar < 1). Negative or NA m/M or roar occurs in not definite situations, such as counts equal to zero for PRE or POST portions. If for one of the conditions there are more than one samples then calculations are performed on average counts.

Usage

         
         
         computeRoars(rds, qwidthTreatment=NA, qwidthControl=NA)
         computeRoars(rds, qwidthTreatment, qwidthControl)
     

Arguments

rds	The RoarDataset or the RoarDatasetMultipleAPA which contains the counts over PRE-POST portions in the two conditions to be compared via roar.
qwidthTreatment	The mean length of the reads in the treatment bam files - used internally for the interaction between RoarDataset and RoarDatasetMultipleAPA objects. The default (NA) calculates this value from the bam and should not be changed.
qwidthControl	The mean length of the reads in the control bam files - used internally for the interaction between RoarDataset and RoarDatasetMultipleAPA objects. The default (NA) calculates this value from the bam and should not be changed.

Value

The RoarDataset or the RoarDatasetMultipleAPA object given as rds with the computeRoars phase of the analysis done. m/M and roars will be held in the RoarDataset object itself in the case of single samples, while in two slots otherwise, but end user normally should not analyze those directly but use totalResults or fpkmResults at the end of the analysis.

Examples

   library(GenomicAlignments)
   gene_id <- c("A_PRE", "A_POST", "B_PRE", "B_POST")
   features <- GRanges(
      seqnames = Rle(c("chr1", "chr1", "chr2", "chr2")),
      strand = strand(rep("+", length(gene_id))),
      ranges = IRanges(
         start=c(1000, 2000, 3000, 3600),
         width=c(1000, 900, 600, 300)),
      DataFrame(gene_id)
   )
   rd1 <- GAlignments("a", seqnames = Rle("chr1"), pos = as.integer(1000), cigar = "300M", strand = strand("+"))
   rd2 <- GAlignments("a", seqnames = Rle("chr1"), pos = as.integer(2000), cigar = "300M", strand = strand("+"))
   rd3 <- GAlignments("a", seqnames = Rle("chr2"), pos = as.integer(3000), cigar = "300M", strand = strand("+"))
   rds <- RoarDataset(list(c(rd1,rd2)), list(rd3), features)
   rds <- countPrePost(rds, FALSE)
   rds <- computeRoars(rds)
    

vodkatad/roar documentation built on March 30, 2020, 2:56 p.m.