collapseBundles: Collapse bundles of transcripts, discard any that represent...

In RamsinghLab/arkas_staging: A package that complements Kallisto for quick, informative *seq analysis

Description

Collapse bundles of transcripts, discard any that represent pointless tests, and optionally prune any whose joined bundle IDs tend to choke downstream packages for e.g. pathway- or network-based enrichment analysis. Note that this function may or may not be optimal for your RNAseq experiment. Please refer to 'Details' for some thought exercises about the nature of 'genes'.

Usage

collapseBundles(kexp, bundleID = "gene_id", read.cutoff = 1,
  discardjoined = TRUE)

Arguments

kexp	A KallistoExperiment (or something very much like it)
bundleID	The column (in mcols(rowRanges(kexp))) of the bundle IDs
read.cutoff	Discard transcripts and bundles with < this many counts
discardjoined	Discard bundles with IDs "joined" by a ";"? (TRUE)

Details

This function sums the estimated counts for each transcript within a bundle of transcripts (where "bundle" is a user-defined identifier, often but not always a 'gene', sometimes a biotype or a class of repeat elements). The default approach is to discard all rows where the maximum count is less than the specified read.cutoff. Since the default cutoff is 1, this means discarding transcripts (and bundles) that were not be detected in any sample. (Filtering tends to increase statistical power at given false-positive rate)

Value

a matrix of summarized counts per sample bundle

See Also

collapseTranscripts

RamsinghLab/arkas_staging documentation built on March 14, 2021, 11:40 a.m.