Expression Methods

allDA

R Documentation

Run many differential abundance/expression methods

Description

Run many differential abundance and expression tests at a time, to easily compare their results

Usage

allDA(
  data,
  predictor,
  paired = NULL,
  covars = NULL,
  tests = c("bay", "ds2", "ds2x", "per", "adx", "znb", "zpo", "msf", "zig", "erq",
    "erq2", "neb", "qpo", "poi", "sam", "lrm", "llm", "llm2", "lma", "lmc", "ere",
    "ere2", "pea", "spe", "wil", "kru", "qua", "fri", "abc", "ttt", "ltt", "ltt2", "tta",
    "ttc", "ttr", "aov", "lao", "lao2", "aoa", "aoc", "vli", "lim", "lli", "lli2", "lia",
    "lic"),
  relative = TRUE,
  cores = (detectCores() - 1),
  p.adj = "fdr",
  args = list(),
  out.all = NULL,
  alpha = 0.1,
  core.check = TRUE,
  verbose = TRUE
)

Arguments

`data`	Either a data.frame with counts/abundances, OR a `phyloseq` object. If a data.frame is provided rows should be taxa/genes/proteins and columns samples, and there should be rownames
`predictor`	The predictor of interest. Either a Factor or Numeric, OR if `data` is a `phyloseq` object the name of the variable in `sample_data(data)` in quotation. If the `predictor` is numeric it will be treated as such in the analyses
`paired`	For paired/blocked experimental designs. Either a Factor with Subject/Block ID for running paired/blocked analysis, OR if `data` is a `phyloseq` object the name of the variable in `sample_data(data)` in quotation.
`covars`	Either a named list with covariates, OR if `data` is a `phyloseq` object a character vector with names of the variables in `sample_data(data)`
`tests`	Character. Which tests to include. Default all
`relative`	Logical. TRUE (default) for compositional data. FALSE for absoloute abundances or pre-normalized data.
`cores`	Integer. Number of cores to use for parallel computing. Default one less than available
`p.adj`	Character. Method for p-value adjustment. See `p.adjust` for details. Default "fdr"
`args`	List. A list with lists of arguments passed to the different methods. See details for more.
`out.all`	If TRUE models will output results and p-values from `anova`/`drop1`. If FALSE will output results for 2. level of the `predictor`. If NULL (default) set as TRUE for multi-class `predictor` and FALSE otherwise
`alpha`	q-value threshold for calling significance. Default 0.1
`core.check`	If TRUE (default) will make an interactive check that the amount of cores specified are desired. Only if `cores>20`. This is to ensure that the function doesn't automatically overloads a server with workers.
`verbose`	If TRUE will print informative messages

Details

mva is excluded by default, as it is slow.

Value

A list of results:

raw - A data.frame with raw p-values from all methods
adj - A data.frame with adjusted p-values from all methods (detection/no-detection from sam)
est - A data.frame with estimates/fold.changes from all relevant methods
details - A dataframe with details from the run
results - A complete list of output from all the methods. Example: Get wilcoxon results from 2. run as such: $results[[2]]["wil"]

Examples

# Creating random count_table and predictor
set.seed(5)
mat <- matrix(rnbinom(500, size = 0.1, mu = 500), nrow = 50, ncol = 10)
pred <- c(rep("Control", 5), rep("Treatment", 5))

# Running allDA to compare methods
# This example uses 1 core (cores = 1). 
# Remove the cores argument to get it as high (and thereby fast) as possible.
res <- allDA(data = mat, predictor = pred, cores = 1)

# View adjusted p-values from all methods
print(res$adj)

# View estimates from all methods
print(res$est)


# Include a paired variable for dependent/blocked samples
subject <- rep(1:5, 2)
res <- allDA(data = mat, predictor = pred, paired = subject, cores = 1)

# Include covariates
covar1 <- rnorm(10)
covar2 <- rep(c("A","B"), 5)
res <- allDA(data = mat, predictor = pred, 
             covars = list(FirstCovar = covar1, CallItWhatYouWant = covar2), cores = 1)

# Data is absolute abundance
res <- allDA(data = mat, predictor = pred, relative = FALSE, cores = 1)

Russel88/DAtest documentation built on March 24, 2022, 3:50 p.m.