computePvals-method: Computes pvalues (Fisher test) on the read counts in this...
In vodkatad/roar: Identify differential APA usage from RNA-seq alignments

Description Usage Arguments Value Examples

This is the third step in the Roar analyses: it applies a Fisher test comparing counts falling on the PRE and POST portion in the treatment and control conditions for every gene. If there are multiple samples for a condition every combinations of comparisons between the samples lists are considered.

1 2	computePvals(rds)

rds

The RoarDataset or the RoarDatasetMultipleAPA which contains the counts over PRE-POST portions in the two conditions to be compared via pvalues.

The RoarDataset or the RoarDatasetMultipleAPA object given as rds with the compute pvalue phase of the analysis done. Pvalues will be held in the RoarDataset object itself in the case of single samples, while in a separate slot otherwise, but end user normally should not analyze those directly but use totalResults or fpkmResults at the end of the analysis.

   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)
   rds <- computePvals(rds)