DESeq2

Question

[t.ru]

Hi,

I am really confused with DESeq2. I only have normal and virus samples. I would like to see upregulated genes in virus samples

dds <-DESeqDataSetFromMatrix(countData = round(count_matrix), colData = col_matrix,
                         design = ~treatment )

I set my reference level by

dds$treatment<-relevel(dds$treatment,ref="normal")

Then I set the constrast after I run the deseq2

contrast1 <- c('treatment',"normal" ,"virus")

I would like to get results for log2foldchange.

comparison_df <-results(dds, contrast = contrast1, alpha = 0.1)

Here, I get the genes I am interested as -log2foldchange values. What I know is they were supposed to be positive.

Could you help me to understand what I am doing wrong? Thank you.

Answer 1

[newDarkAB09834]

In DESeq2, condition treated vs untreated is calculated as log2(treated/untreated) so,

res <- results(dds, contrast=c("condition", "treated", "untreated"))

So I think in your case it should be:

contrast=c("treatment", "virus", "normal")

it the normal level is your reference level.

[swbarnes2]

Yes. When you specify the contrast, it doesn't matter what you set as the reference with relevel.