Question: DESeq2 Control for differences between patients
gravatar for gkuffel
9 months ago by
Loyola University
gkuffel10 wrote:


I have a relatively simple RNA-Seq experiment. I have 10 samples in my control group and 10 samples in my treatment group. Each sample is from a different patient. How can I test for differences between control and treatment groups but control for the natural differences we might see between patients. Your help is greatly appreciated.

ADD COMMENTlink modified 9 months ago by swbarnes250 • written 9 months ago by gkuffel10
gravatar for swbarnes2
9 months ago by
swbarnes250 wrote:

You have 10 control samples and 10 treatment samples.  If you wanted to control for sex, or age (maybe by splitting the samples into a few age groupings) you could tell it that information, and make use of that, but it's already going to treat your samples as if they were biological replicates.

ADD COMMENTlink written 9 months ago by swbarnes250

Thank you for the response. So I do not see any differential expression when using design =~ condition and p-adj < 0.1. Is there anything else I can modify or try? Also, if I look at p values instead of p-adj then I see some differential expression (about 1000 genes).

ADD REPLYlink written 9 months ago by gkuffel10

Unadjusted p-values are not the right thing to be looking at.  Stick with adjusted p-values.  As always, XKCD explains why:

ADD REPLYlink written 9 months ago by swbarnes250

Take a look at the MA plot (see vignette). This will give you a senses of the differences across treatment. You can also look at the plotCounts() function to see why the top gene doesn't have an adjusted p-value less than 0.1. This typically means small differences across group relative to the variance within group.

ADD REPLYlink written 9 months ago by Michael Love20k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 370 users visited in the last hour