Hi

I'm trying to use DESeq2 to find differentially expressed genes between sample with replicates and sample without replicates. I have 2 condition, control and tumour. I am new to Deseq2 package, do you think my command lines are OK for my experiment?

Thanks in advance

code for find DE in tumour vs control (tumour /control )

> countsTable <- read.delim ("/feature/Cont_control_34.txt", header=TRUE, row.names=1)

> pdata = data.frame(condition = factor(c("control","tumour","tumour","tumour")))

> dds <- DESeqDataSetFromMatrix(countData=countsTable, colData = pdata, design=~condition)

> colData(dds)$condition <- relevel(colData(dds)$condition, "control","tumour")

> design(dds)

~condition

> dds <- DESeq(dds)

> res <- results(dds)

> res <- res[order(res$padj),]

> sig.up.results <- res[which(res$padj < 0.05 & res$log2FoldChange > 1),]

> sig.down.results <- res[which(res$padj < 0.05 & res$log2FoldChange < -1),]

> sig.results <- res[which(res$padj < 0.05),]

> sig.results <- sig.results[order(sig.results$log2FoldChange, decreasing=TRUE),]

Yes that looks correct.

(Except you might want to know that relevel() only takes the reference level as the second argument, you don't provide it with the other levels. It just ignores the other arguments you provided. See ?relevel.)

There was recently another similar question here about when one condition has no replicates

Can I run DESeq2, if I have repeats for my control, but no repeats for my test condition?

(I moved your post from an Answer to a Comment.)

Take a look at what you see in the help page for ?relevel:

Usage

relevel(x, ref, ...)

Arguments

x    
an unordered factor.

ref    
the reference level, typically a string.

...    
additional arguments for future methods.

So this means that anything after the second argument is ignored by the function.

You can also convince yourself that extra arguments to relevel have no function by playing around with some R code yourself, and seeing that whatever you put after the second argument doesn't change the order of levels for the factor:

> x <- factor(c("a","b","c"))
> x
[1] a b c
Levels: a b c
> relevel(x, "b")
[1] a b c
Levels: b a c
> relevel(x, "b", "c")
[1] a b c
Levels: b a c
> relevel(x, "b", 42)
[1] a b c
Levels: b a c