mod.t.test: Moderated t-Test

In MKmisc: Miscellaneous Functions from M. Kohl

Moderated t-Test

Description

Performs moderated t-tests based on Bioconductor package limma.

Usage

mod.t.test(x, group = NULL, paired = FALSE, adjust.method = "BH",
           sort.by = "none")

Arguments

x	a (non-empty) numeric matrix of data values.
group	an optional factor representing the groups.
paired	a logical indicating whether you want a paired test.
adjust.method	see p.adjust
sort.by	see toptable

, where "logFC" corresponds to difference in means.

Details

The function uses Bioconductor package limma to compute moderated t-tests. For more details we refer to ebayes.

Value

A data.frame with the results.

References

B. Phipson, S. Lee, I.J. Majewski, W.S. Alexander, G.H. Smyth (2016). Robust hyperparameter estimation protects against hypervariable genes and improves power to detect differential expression. Annals of Applied Statistics 10(2), 946-963.

See Also

t.test

Examples

## One-sample test
X <- matrix(rnorm(10*20, mean = 1), nrow = 10, ncol = 20)

mod.t.test(X)
## corresponds to
library(limma)
design <- matrix(1, nrow = ncol(X), ncol = 1)
colnames(design) <- "A"
fit1 <- lmFit(X, design)
fit2 <- eBayes(fit1)
topTable(fit2, coef = 1, number = Inf, confint = TRUE, sort.by = "none")[,-4]

## Two-sample test
set.seed(123)
X <- rbind(matrix(rnorm(5*20), nrow = 5, ncol = 20),
           matrix(rnorm(5*20, mean = 1), nrow = 5, ncol = 20))
g2 <- factor(c(rep("group 1", 10), rep("group 2", 10)))

mod.t.test(X, group = g2)
## corresponds to
design <- model.matrix(~ 0 + g2)
colnames(design) <- c("group1", "group2")
fit1 <- lmFit(X, design)
cont.matrix <- makeContrasts(group1vsgroup2="group1-group2", levels=design)
fit2 <- contrasts.fit(fit1, cont.matrix)
fit3 <- eBayes(fit2)
topTable(fit3, coef = 1, number = Inf, confint = TRUE, sort.by = "none")[,-4]

## Paired two-sample test
mod.t.test(X, group = g2, paired = TRUE)

MKmisc documentation built on Nov. 20, 2022, 1:05 a.m.