Question: DESeq2: Small treatment group, large control group
0
3 months ago by
kel.chee860 wrote:

Hi I would to use DESeq2 to analyse NanoString miRNA expression data. I have a group of 5 - 6 conditioned group, and large set of controls available for comparison (>96). In this case I'm looking at different expression between two cancer types. If I include all cases in one analysis, how reliable would the results be? I also think about  breaking the controls into groups of 5-6 cases and perform multiple analysis to have equal size of conditioned and control groups. Would this work? Thanks!

deseq2 sample size • 125 views
modified 3 months ago by Michael Love22k • written 3 months ago by kel.chee860
Answer: DESeq2: Small treatment group, large control group
0
3 months ago by
Michael Love22k
United States
Michael Love22k wrote:

This question was asked very recently on the site, but it’s hard to find similar questions. My answer was that it’s not technically a problem although the more samples in the small group the better obviously.

Searching through google with a site:support.bioconductor.org modifier(?) is almost always better than using the search functionality baked into this board. For instance, googling for: unbalanced groups rnaseq site:support.bioconductor.org gets you a few hits:

They all speak to this problem ... maybe the first one was the one you were thinking of?

I think it’s not in that list... it’s tricky to even think of the possible variations on keywords to find it. Not an easy solution to find related content on forums!

I found it just by scrolling through DESeq2 posts:

Question regarding class imbalance and low sample size for DESeq2