Skip to contents

Simulate SummarizedExperiment tests

Source: R/jam_simulate_se_test.R
simulate_se_test.Rd

Simulate SummarizedExperiment tests, with adjustable batch effect.

Usage

simulate_se_test(
  ngroups = 4,
  nreps = 5,
  nrow = 250,
  multiplier = 1.5,
  offset = 7,
  batch_ex = 0.15,
  hit_exp = 2,
  hit_fraction = 1/2,
  hit_max = 2.8,
  noise_factor = 1,
  seed = 123,
  assay_name = "counts",
  sparsity = 0,
  verbose = FALSE,
  ...
)

Arguments

ngroups

integer number of experimental groups

nreps

integer number of replicates per group, can be used to provide the number of replicates for each group in order.

nrow

integer number of rows (measurements)

multiplier

numeric value multiplied by rnorm() to adjust the magnitude of values produced, default 1.

offset

numeric value added to the output of rnorm(), default 7.

batch_ex

numeric adjustment for batch effect, default 0.15 is a moderate batch effect, it would require adjustment or blocking or covariate for effective analysis.

hit_exp

numeric default 2, adjust the "sharpness" of differential changes, higher values are sharper, with fewer strong hits.

hit_fraction

numeric value between 0 and 1 indicating the fraction of rows to simulate as having a fold change, default 1/2.

hit_max

numeric maximum value for a simulated fold change, default 2.8, intended to be interpreted as log2 of a 7-fold change.

noise_factor

numeric multiplied by rnorm() to add additional noise.

seed

numeric passed to set.seed() when provided, for reproducible random output.

assay_name

character name to use for the assay name in the output.

sparsity

numeric value from 0 to 1, default 0, indicating the fraction of values that are converted to NA to simulate sparse data measurements. It can be provided as a vector and applied to each group in order. In some proteomics datasets, control samples may be substantially more sparse than other groups, for example if the control is non-targeted IP, or negative control. In this case, data can be simulated by using sparsity=c(0.6, 0, 0) for three groups.

verbose

logical indicating whether to print verbose output.

...

additional arguments are ignored.

Value

SummarizedExperiment object

Details

This function extends make_se_test() in two ways:

  1. For four-group scenario, it will generate two-factor changes.

  2. Batch effects are generated, adjustable with batch_ex. Use batch_ex=0 for no batch effect. See examples for demonstration of centering within batch as visual cue that batch effect is well-defined.

See also

Other jamses SE utilities: geomx_to_se(), make_se_test(), se_collapse_by_column(), se_collapse_by_row(), se_detected_rows(), se_normalize(), se_rbind(), se_to_assay_data(), se_to_assay_names(), se_to_rowcoldata()

Examples

# batch effects
seb <- simulate_se_test();
#> Warning: first element used of 'each' argument
#> Warning: first element used of 'each' argument
#> Warning: data length [1249] is not a sub-multiple or multiple of the number of rows [250]
#> Warning: first element used of 'length.out' argument
#> Warning: first element used of 'length.out' argument
#> Warning: first element used of 'length.out' argument
#> Warning: first element used of 'length.out' argument
hmb <- heatmap_se(seb,
   controlSamples=colnames(seb)[c(1:5, 11:15)],
   column_title="With batch effect\nglobal-centered")
hmb


# when centering versus a sparse control group, some values can be lost:
hmbg <- heatmap_se(seb,
   controlSamples=colnames(seb)[c(1:5, 11:15)],
   centerby_colnames="batch",
   column_title="With batch effect\ncentered by batch")
hmb + hmbg
#> Warning: Heatmap/annotation names are duplicated: centered expression


hmbc <- heatmap_se(seb,
   correlation=TRUE,
   controlSamples=colnames(seb)[c(1:5, 11:15)],
   column_title="With batch effect\nglobal-centered")
hmbc


hmbgc <- heatmap_se(seb,
   correlation=TRUE,
   controlSamples=colnames(seb)[c(1:5, 11:15)],
   centerby_colnames="batch",
   column_title="With batch effect\ncentered by batch")
hmbgc

hmbc + hmbgc
#> Warning: Heatmap/annotation names are duplicated: correlation of centered
#> expression