correct_gc_bias: correct_gc_bias: Estimate and correct GC bias in coverage

Description

This generic function is used to run to calculate and correct GC-content-based coverage bias

This function optionally estimates and then corrects the GC bias based on a GC-content factor file that is either generated or provided by the user using a sliding window approach. It creates a GC factor file and a corrected coverage file, both of which are placed in the output directory under `/gc`.

Usage

correct_gc_bias(obj, ...)
# S3 method for default
correct_gc_bias(
  obj = NULL,
  df = NULL,
  sample.names = NULL,
  ano.file,
  out.dir = NULL,
  ncpus = 1,
  file.raw.coverage = NULL,
  file.corrected.coverage = NULL,
  file.gc.factor = NULL,
  win.size = 75,
  gc.step = 0.01,
  estimate_gc = TRUE,
  verbose = FALSE,
  ...
)
# S3 method for RCNA_object
correct_gc_bias(obj, verbose = FALSE, ...)

Value

A RCNA_analysis class object that describes the input parameters and output files generated by this step of the workflow.

Arguments

obj: A RCNA_object type object - parameters will be pulled from the object instead, specifically from the `gcParams` slot.
...: Additional arguments (unused)
df: Path to the config file, or a `data.frame` object containing the valid parameters. Valid column names are `file.raw.coverage`, `file.gc.factor`, `file.corrected.coverage`, and `sample.names`. Additional columns will be ignored.
sample.names: Character vector of sample names. Alternatively can be specified in `df`.
ano.file: Location of the annotation file. This file must be in CSV format and contain the following information (with column headers as specified): "feature,chromosome,start,end".
out.dir: Output directory for results. A subdirectory for results will be created under this + `/nkr/`.
ncpus: Integer number of CPUs to use. Specifying more than one allows this function to be parallelized by feature.
file.raw.coverage: Character vector listing the raw input coverage files. Must be the same length as `sample.names`. Alternatively can be specified in `df`.
file.corrected.coverage: Character vector listing the corrected input coverage files. If not specified new names will be generated based on the raw coverage files.
file.gc.factor: Character vector listing the GC factor files used to correct coverage. If `estimate_gc=FALSE` then this must be provided. Otherwise it is ignored.
win.size: Size in base pairs of the sliding window used to estimate and correct the GC bias.
gc.step: Bin size for GC bias in the GC factor file. If the GC factor file is provided then the file must have corresponding bin sizes.
estimate_gc: Logical determining if GC content estimation should be performed. If set to `FALSE` then a factor file must be provided via `file.gc.factor` or in `df`.
verbose: If set to TRUE will display more detail

Details

This function can be run as a stand-alone or as part of run_RCNA

The `df` argument corresponds to the `gcParams` matrix on RCNA_object. Valid column names are `sample.names`, `file.raw.coverage`, `file.corrected.coverage`, and `file.gc.factor`. The `file.gc.factor` column is not required if `estimate_gc=TRUE`. Additional columns will be ignored.

For more parameter information, see estimate_nkr.default.

Examples

Run this code

## Run GC-bias estimation and correction on example object
# See \link{example_obj} for more information on example
example_obj@ano.file <- system.file("examples" ,"annotations-example.csv",
 package = "RCNA")
raw.cov <- system.file("examples", "coverage",
                       paste0(example_obj@sample.names, ".txt.gz"), package = "RCNA")
example_obj@gcParams$file.raw.coverage <- raw.cov
example_obj
# Create output directory
dir.create(file.path("output", "gc"), recursive = TRUE)
# Estimate and correct GC bias, append results
correct_gc_analysisObj <- correct_gc_bias(example_obj)
example_obj@commands <- c(example_obj@commands, correct_gc_analysisObj)
system("rm -rf output")