Question about contrasts in DESeq2
1
1
Entering edit mode
marinaw ▴ 10
@9c8b15cf
Last seen 7 weeks ago
Canada

Hi all,

I'm using interaction terms for DGE analysis I'm doing using DESeq2. I have a case and control group (SA and CTRL, respectively) and males and females per group (Male, Female). I would like investigate sex differences between cases and controls. Here is the design formula:

dds <- DESeqDataSetFromMatrix(countData = countData, colData = colData, design = ~  Age_scale + pH_scale + Sex + Group + Sex:Group

However, I'm having trouble truly understanding what goes on when obtaining results in the presence/absence of using contrast=list(c(.

To explain myself more clearly, what is the difference between:

1) Using no specified contrast when obtaining results. I understand that DESeq2 takes by default the last term, so, you have the results of the interaction effect alone. It is my understanding that what is happening here is: group effect in the female group - the group effect in the male group (group effect is SA vs CTRL). (Male and CTRL as references).

2) Using for example contrast=list( c("Group_SA_vs_CTRL", "SexFemale.GroupSA"))) when obtaining results.

In my mind, 2 is identical to 1. But is that actually true? Can someone explain to me the precise steps/differences in how results are determined, if there are any?

Thanks!

DESeq2 • 247 views
ADD COMMENT
0
Entering edit mode
@mikelove
Last seen 14 minutes ago
United States

There is a section on interactions in the vignette, but if you want to know more, I recommend to work with someone familiar with linear models in R.

Also ?results has a lot of information about how the tables can be printed, e.g. the different ways to use contrast.

ADD COMMENT
0
Entering edit mode

Thank you Michael!

ADD REPLY

Login before adding your answer.

Traffic: 452 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6