polyester (version 1.6.0)

add_gc_bias: add GC bias to a count matrix

Description

Given a matrix with rows corresponding to transcripts and sample-specific GC bias models, bias the count matrix using the bias model.

Usage

add_gc_bias(readmat, gcbias, transcripts)

Arguments

readmat

matrix of counts, with rows corresponding to features (transcripts) and columns corresponding to replicates

gcbias

List of GC bias models to add to readmat. Must have length equal to the number of columns of readmat. List elements must either be integers 0 through 7, where 0 means no bias and 1-7 correspond to built-in GC bias models, or objects of class loess which can predict a deviation from overall mean count (on the log scale) given a GC percentage between 0 and 1.

transcripts

DNAStringSet object containing the sequences of the features (transcripts) corresponding to the rows of readmat. Length must be equal to the number of rows in readmat.

Value

matrix of the same size as readmat, but with counts for each replicate biased according to gcbias.

Details

Designed for internal use in simulate_experiment functions.

Examples

library(Biostrings)
  fastapath = system.file("extdata", "chr22.fa", package="polyester")
  numtx = count_transcripts(fastapath)
  transcripts = readDNAStringSet(fastapath)

  # create a count matrix:
  readmat = matrix(20, ncol=10, nrow=numtx)
  readmat[1:30, 1:5] = 40

  # add biases randomly: use built-in bias models
  set.seed(137)
  biases = sample(0:7, 10, replace=TRUE)
  readmat_biased = add_gc_bias(readmat, as.list(biases), transcripts)