normalize_expt: Normalize the data of an expt object. Save the original...

In elsayed-lab/hpgltools: A pile of (hopefully) useful R functions

Normalize the data of an expt object. Save the original data, and note what was done.

Description

It is the responsibility of normalize_expt() to perform any arbitrary normalizations desired as well as to ensure that the data integrity is maintained. In order to do this, it writes the actions performed in expt$state and saves the intermediate steps of the normalization in expt$intermediate_counts. Furthermore, it should tell you every step of the normalization process, from count filtering, to normalization, conversion, transformation, and batch correction.

Usage

normalize_expt(
  expt,
  transform = "raw",
  norm = "raw",
  convert = "raw",
  batch = "raw",
  filter = FALSE,
  annotations = NULL,
  fasta = NULL,
  entry_type = "gene",
  use_original = FALSE,
  batch1 = "batch",
  batch2 = NULL,
  batch_step = 4,
  low_to_zero = TRUE,
  thresh = 2,
  min_samples = 2,
  p = 0.01,
  A = 1,
  k = 1,
  cv_min = 0.01,
  cv_max = 1000,
  na_to_zero = FALSE,
  adjust_method = "ruv",
  verbose = TRUE,
  ...
)

Arguments

expt	Original expt.
transform	Transformation desired, usually log2.
norm	How to normalize the data? (raw, quant, sf, upperquartile, tmm, rle)
convert	Conversion to perform? (raw, cpm, rpkm, cp_seq_m)
batch	Batch effect removal tool to use? (limma sva fsva ruv etc)
filter	Filter out low/undesired features? (cbcb, pofa, kofa, others?)
annotations	Used for rpkm – probably not needed as this is in fData now.
fasta	Fasta file for cp_seq_m counting of oligos.
entry_type	For getting genelengths by feature type (rpkm or cp_seq_m).
use_original	Use the backup data in the expt class?
batch1	Experimental factor to extract first.
batch2	Second factor to remove (only with limma's removebatcheffect()).
batch_step	From step 1-5, when should batch correction be applied?
low_to_zero	When log transforming, change low numbers (< 0) to 0 to avoid NaN?
thresh	Used by cbcb_lowfilter().
min_samples	Also used by cbcb_lowfilter().
p	Used by genefilter's pofa().
A	Also used by genefilter's pofa().
k	Used by genefilter's kofa().
cv_min	Used by genefilter's cv().
cv_max	Also used by genefilter's cv().
na_to_zero	Sometimes rpkm gives some NA values for very low numbers.
adjust_method	Given a set of sv estimates, change the counts with this method.
verbose	Print what is happening while the normalization is performed? I am not sure why, but I think they should be 0.
...	more options

Value

Expt object with normalized data and the original data saved as 'original_expressionset'

See Also

[convert_counts()] [normalize_counts()] [batch_counts()] [filter_counts()] [transform_counts()]

Examples

## Not run: 
 normed <- normalize_expt(exp, transform='log2', norm='rle', convert='cpm',
                          batch='raw', filter='pofa')
 normed_batch <- normalize_expt(exp, transform='log2', norm='rle', convert='cpm',
                                batch='sva', filter='pofa')

## End(Not run)

elsayed-lab/hpgltools documentation built on May 9, 2024, 5:02 a.m.