Contrasts in DESeq2
1
0
Entering edit mode
@bahmanikmsuedu-23146
Last seen 22 months ago
Michigan State University

Hi, I'm using DESeq2 on 3 high sugar genotypes and 3 low sugar genotypes. None of the genotypes or sugar groups are reference, but DESeq2 considers one of the genotypes as reference (I think just based on order of the numbers assigned to each genotype). Since I just want to find DEGs between low and high sugar genotypes, I did my desired contrasts (each low sugar genotype against each high sugar genotype), and found DEGs. I was wondering if what I did, regarding the reference assumption by DESeq2 and my contrasts, is correct, and if it is not, what do I need to do? Thank you,

deseq2 contrast • 1.1k views
1
Entering edit mode
@mikelove
Last seen 2 days ago
United States

With contrast in results() you can compare any two groups you want. You don't have to worry which is reference if you specify contrast.

0
Entering edit mode

After finding DEGs in a contrast between one high and one low sugar genotype, genes in which one of the 2 genotypes are up or down regulated in comparison to the other one? Thank you,

0
Entering edit mode

results(dds, contrast=c("condition","C","B")) meaning genes with logFC > 0 are overexpressed in C.

0
Entering edit mode

Genes with logFC > 0 are overexpressed in C (the first genotype written in the contrast), and genes with logFC < 0 are underexpressed in B (the second genotype written in the contrast); is this correct? Then how do we find overexpressed genes in B, and also underexpressed in C? Thank you,

1
Entering edit mode

No, logFC < 0 means underexpressed in C and overexpressed in B. The fold change direction is interchangable. C vs B is the same towards the stats as B vs C, only the direction changes.

0
Entering edit mode

I see now. Thanks a lot for the clarification.

0
Entering edit mode

Hi Micheal,

I'm using contrast this way:

res <- results( dds, contrast=c("NAS","NAR","NNAR" ))

and when I try:

res <- results( dds, contrast=c("NAS","NNAR","NAR" ))

I still get the same sign for the my logfold change while it should change because I'm changing the alternative and reference. Do you have any idea why its not working?

Thanks !

0
Entering edit mode

Double check your code, maybe give these tables different names.

0
Entering edit mode

Perhaps your condition was set as factor before you used Deseq2