DESEQ2 study design for different time points between controls and high dose samples.

Question

[salman_96]

Hi, My study consists of exposure of a drug. There are approximately 25 controls and 30 cases exposed to the drug. I have prepared a dummy dataset that would help me put my question here. The Metadata looks like this:

time_days<-(factor(c(10,10,10,10,10,10,4,4,4,4,4,4,1,1,1,1,1,1)))
coldata <- data.frame(dose,
                  condition = factor(c(
                    rep("control", 3),
                    rep("high", 3),
                    rep("control", 3),
                    rep("high", 3),
                    rep("control", 3),
                    rep("high", 3))),time_days)
coldata$Groups_of_Interest <- paste(coldata$condition,coldata$time_days,sep = "_")
coldata

The cases and controls are divided into different groups based on days(time days column). I am interested in the column Groups_of_interest (see below image)

Which one of the following study designs do I use to make comparisons between all groups.

dds <- DESeqDataSetFromMatrix( countData = x, colData = coldata,
design= ~ condition + time_days)

library(DESeq2)

 dds <- DESeqDataSetFromMatrix(   countData = x,  
> colData = coldata,   design= ~ Groups_of_Interest)

I am interested in:

• Groups_of_Interest_control_10_vs_control_1, Groups_of_Interest_control_4_vs_control_1, Groups_of_Interest_high_1_vs_control_1, Groups_of_Interest_high_10_vs_control_1, Groups_of_Interest_high_4_vs_control_1

Running design= ~ Groups_of_Interest does provide this information when I check resultnames(dds) but is it the correct way?

Thank you!

[ATpoint]

Yes, that so-called "full factorial" design would be the easiest while catching all the experimental factors. Check contrasts in the vignette, with that you can build any comparison you like.

[salman_96]

Hi, Thank you for responding. The column Groups of interests contains information about both (other columns) condition and time_days. Would it be sufficient to use only this column in the design ~Groups_of_Interest

Thank you!

[salman_96]

I would only like to see the comparison between the same time groups in my groups of interest for example: controls at 10 days vs high exposure at 10 days, controls at 4 days vs high exposure at 4 days.

Thankyou!

[Basti]

Yes you can choose the comparison you want with the contrasts argument of the results function as ATpoint indicated. Have a look at this to see how to specify your contrasts : http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#contrasts