standardFilter {roar} | R Documentation |
RoarDataset
object
or a RoarDatasetMultipleAPA
objectThe last step of a classical Roar analyses: it returns a dataframe containing m/M values, roar values, pvalues and estimates of expression (a measure recalling FPKM). Only the genes with an expression estimate bigger than a given cutoff will be considered.
standardFilter(rds, fpkmCutoff)
rds |
The |
fpkmCutoff |
The cutoff that will be used to determine if a gene is expressed or not. |
For RoarDataset
and RoarDatasetMultipleAPA
:
The resulting dataframe will be identical to that returned by fpkmResults
but
it will contains rows relative only with genes with an expression estimate (treatment or controlValue)
bigger than the given fpkmCutoff in both the conditions and with sensitive m/M and roar values (it
removes negative or NA
m/M values/roar - these values arise when there aren't enough information to draw a conclusion
about the shortening/lengthening of the gene).
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) dat <- standardFilter(rds, 1)