mirl: Multiple Iterations of Rarefying Libraries

In escamero/mirlyn: Multiple Iterations of Rarefying for Library Normalization

Multiple Iterations of Rarefying Libraries

Description

mirl() will repeatedly rarefy to a user defined library size, for a specified number of replications to characterize the variation introduced through random subsampling. The implementation of mirl() allows for rarefying to be a statistically valid library size normalization technique for diversity analyses. Users are encouraged to conduct exploratory analysis to identify optimal rarefied library sizes for their data that minimizes variation in samples. Users may have to choose between rarefying to smaller than preferred library sizes or discarding small library size samples but the implementation of mirl() generally allows for rarefying to small library sizes at the loss of resolution and precision in results.

Usage

mirl(
  x,
  libsize = min(phyloseq::sample_sums(x)),
  rep = 1000,
  set.seed = NULL,
  trimOTUs = FALSE,
  replace = FALSE,
  mc.cores = 1L
)

Arguments

x	The phyloseq object
libsize	The specified library size to rarefy to. By default, mirl() will rarefy to the minimum library size found in samples.
rep	The number of times rarefying will be repeated. By default, mirl() will repeatedly rarefy 1000 times.
set.seed	The seed value for reproducibility.
trimOTUs	A Boolean value determining whether OTUs that are absent from all samples after rarefying should be removed.
replace	A Boolean value determining whether subsampling should be performed with or without replacement. See Cameron et al., 2020 for further details.
mc.cores	From parallel::mclapply().

Value

A list of phyloseq objects.

Examples

library(mirlyn)
data(example)

## Not run: 
mirl_object <- mirl(example, libsize = 10000, rep = 100, set.seed=120)

## End(Not run)

escamero/mirlyn documentation built on Aug. 6, 2023, 9:30 p.m.