This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Understanding multifactor designs and control variables in DESeq2

I've read the DESeq2 documentation but I still don't understand how to analyze, interpret, and visualize multifactor designs for RNA-seq. Below is an example of what I am trying to figure out

Here, I have 4 groups that are differentiated by their values of controlVar and tx. I am interested in:

  1. How groups 2, 3, and 4 differ from group 1
  2. How groups 3 and 4 differ from each other
  3. How groups 3 and 4 differ from group 2

Any help on how to write the design formula to accomplish these would be tremendously appreciated. If I can provide further information, I will do so.

Here was my thought process:

  1. I'm guessing this looks at the effect of tx controlling for (i.e. stratifying by) controlVar? design = ~ controlVar + tx
  2. I don't think I am interested in the interaction of controlVar and tx because in theory the controlVar should not be affecting anything
  3. Would tx for groups 3 and 4 be considered a nested variable? If so, I don't know how to write this.

enter image description here

rna-seq deseq2
• 1,459 views • link
written by kldsajf;ldsj;l • kldsajf;ldsj;l

1 answer

https://bioinformatics-core-shared-training.github.io/cruk-summer-school-2021/RNAseq/Markdowns/10_DE_analysis_with_DESeq2.Solutions.html

This page explains how to test which which model is best for your data. It is also explained in the documentation in the "Likelihood ratio test" section. Make sure to set factor levels correctly.

• 0 views • link
written by FrankStarling • FrankStarling

Thank you, but this does not help me with the nested variables issue. Basically, samples 4-12 have tx values that are nested within contralVar, and I'm not sure how to analyze this properly.

• 0 views • link
written by kldsajf;ldsj;l • kldsajf;ldsj;l

Log in to answer this question.

More posts like this