How to perform paired t-test in R for obtaining p-value of each gene?

Question

[prithvi.mastermind]

I'm interested in paired t-test for obtaining p-values for each gene. I used the t.test function in R, but it only gives one single p-value. How do I get p-value for all genes? I have 46 pairs of tumor and normal samples.

####Load necessary packages######

library(NOISeq)
library(edgeR)
library(DESeq2)
library(stats)
library(limma)
library(Rsamtools)

####Load the data #####

LIMS <- read.delim(file.choose(), row.names=1)

#####Perform paired t-test between tumor and normal samples####

Tumor <- unlist(LIMS[, 1:46])
Normal <- unlist(LIMS[, 47:92])
y <- t.test(Tumor, Normal, paired = TRUE)

[Jeremy]

I think you would need to run a 2-way ANOVA to get p-values for all genes. However, it's my understanding that false discovery rate (FDR) is the gold standard for RNA-seq.

[Marco Pannone]

Following the DESeq2 pipeline, you directly get the adjusted p-values for all genes.

[prithvi.mastermind]

Actually all my samples are paired and hence I want to apply paired t-test. I guess DESeq2 only offers unpaired two-sample t-test and not paired t-test.

[Trivas]

Want to make sure you saw ATpoint's post. You would include that information into the design by adding another column to your colData

Answer 1

[ATpoint]

You have to include the pairing information into the design, then you get a proper paired test in DESeq2. It’s a FAQ, see http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#can-i-use-deseq2-to-analyze-paired-samples

Answer 2

[Gordon Smyth]

You already computed genewise paired t-tests using limma a few days ago: How to apply two tailed paired t-test in limma for DEGs selection?