createSplits: createSplits

In mcalgaro93/benchdamic: Benchmark of differential abundance methods on microbiome data

createSplits

Description

Given a phyloseq or TreeSummarizedExperiment object from which the random splits should be created, this function produces a list of 2 data.frame objects: Subset1 and Subset2 with as many rows as the number of splits and as many columns as the half of the number of samples.

Usage

createSplits(
  object,
  assay_name = "counts",
  varName = NULL,
  paired = NULL,
  balanced = TRUE,
  N = 1000
)

Arguments

object	a phyloseq or TreeSummarizedExperiment object.
assay_name	the name of the assay to extract from the TreeSummarizedExperiment object (default assayName = "counts"). Not used if the input object is a phyloseq.
varName	name of a factor variable of interest.
paired	name of the unique subject identifier variable. If specified, paired samples will remain in the same split. (default = NULL).
balanced	If TRUE a balanced design will be created for the splits (default balanced = TRUE).
N	number of splits to generate.

Value

A list of 2 data.frame objects: Subset1 and Subset2 containing N rows and half of the total number of samples columns. Each cell contains a unique sample identifier.

Examples

data(ps_plaque_16S)
set.seed(123)

# Balanced design for repeated measures

# Balanced design for independent samples
splits_df <- createSplits(
    object = ps_plaque_16S, varName =
        "HMP_BODY_SUBSITE", balanced = TRUE, N = 100
)

# Unbalanced design
splits_df <- createSplits(
    object = ps_plaque_16S, varName =
        "HMP_BODY_SUBSITE", balanced = FALSE, N = 100
)

mcalgaro93/benchdamic documentation built on Sept. 26, 2024, 6:34 p.m.