limma_wrapper: Perform linear regression with limma
In ftwkoopmans/msdap: Mass Spectrometry Downstream Analysis Pipeline

View source: R/stats_differential_abundance.R

limma_wrapper

R Documentation

Perform linear regression with limma

Description

Please refer to the vignette on custom limma models at the MS-DAP GitHub repository for a walk-through with examples.

Usage

limma_wrapper(
  x = protein_data,
  npep = NULL,
  model_matrix,
  contrasts,
  limma_block_variable = NULL,
  ebayes_trend = FALSE,
  ebayes_robust = FALSE,
  deqms = TRUE,
  return_table = TRUE
)

Arguments

`x`	result from `get_protein_matrix(... , include_npep = TRUE)`. Alternatively, a protein log2 intensity matrix (in this case, also provide the `npep` parameter if you want to apply DEqMS)
`npep`	optionally, the number of peptides-per-protein for each row in input protein matrix. Only needed when providing a matrix for parameter `x` (i.e. you can leave this NULL when `x` is output from `get_protein_matrix(... , include_npep = TRUE)`)
`model_matrix`	result from `stats::model.matrix()`
`contrasts`	contrasts defined with `limma::makeContrasts()` that should be extracted from the linear regression model
`limma_block_variable`	option passed to `limma::lmFit()` to fit block design. To skip (default), set NULL. Otherwise, this should be a character vector of the same length as nrow(model_matrix) so it describes a block/group per respective sample in the model/design matrix
`ebayes_trend`	option passed to `limma::eBayes()`. Defaults to `FALSE` (same as limma default)
`ebayes_robust`	option passed to `limma::eBayes()`. Defaults to `FALSE` (same as limma default)
`deqms`	apply DEqMS? Defaults to `TRUE`. If `FALSE`, returns `limma::eBayes()` result as-is
`return_table`	if `TRUE`, the default, returns a table with protein_id, log2fc, pvalue, adjusted pvalue, contrast (i.e. this is the output from MS-DAP function `limma_fit_extract_stats()`). If `FALSE`, returns the limma fit object as-is