Question

DESeq2 on 2 genotypes versus 1 control

0

Entering edit mode

cardin.julie • 0

@cardinjulie-12735

Last seen 6.9 years ago

Hi,

I am student in bioinformatics. I had read the nice vignette for DESeq2. There is a lot of nice information. But even if it is great, I still have a question to address and I am not sure which approach I should use. I have already done DESeq2 for one group versus control and it works well. I was reading about conditions... and I am not sure how to treat it when I have 2 different genotypes and I want what they have in Common compare to a control group. Any suggestion?

Thank you very much

deseq2 • 938 views

ADD COMMENT • link updated 7.5 years ago by Michael Love 42k • written 7.5 years ago by cardin.julie • 0

0

Entering edit mode

If, by common, you mean having an identical difference (upto noise) in both genotypes when compared to control, then you could create another variable along the lines of condition2 <- ifelse(condition=="control", "control", "common_genotype") which would 'pool' the two genotypes, but this is probably artificial and what you want is Michael's answer.

ADD REPLY • link 7.5 years ago Gavin Kelly ▴ 690

score 0 · Answer 1 · 2017-06-04

There's not a single contrast that you can get the set of genes which are DE in group 1 vs control and group 2 vs control. I'd suggest you build the two results tables and then either take the FDR bounded subset in both groups (this final set doesn't have an FDR interpretation though), or you could combine the p-values for the two groups and then run p.adjust(res$pvalue, method="BH").