0
$\begingroup$

I have a figure in a paper where I show the log2(fold-change) (obtained with DESeq2) between two groups (based on genotype) for genes within a specific hallmark gene set, and a reviewer is asking whether the differences are statistically significant.

I have not been able to find any methods for comparing expression of groups of genes like this, and i'm sure there is a method or a reason why it isn't done that I'm not aware of.

To be extra clear, what i have is two sets of log2FC values for a set of genes (one for one genotype (drug-treated vs. untreated) and the same for a second genotype), and I'm looking for a statistical comparison of these two groups of Log2FC measurements.

Thanks

Improve this question
$\endgroup$
2
  • $\begingroup$ They want you to find genes whose response to the drug is different based on the genotype? If so, you can add an interaction term to your DESeq2 regression formula ~ condition + genotype + condition:genotype. Refer to the DESeq2 documentation and help("results", package="DESeq2") for more information on multifactor design and interaction terms. $\endgroup$
    – rpolicastro
    Commented Sep 28, 2020 at 14:26
  • $\begingroup$ Okay yeah that makes sense for how to obtain differences in the response, but then how might I statistically compare these differences in response for a group of genes (I.e a gene set)? Would you just do something simple like a t test / Wilcox test? $\endgroup$
    – Reilstein
    Commented Sep 28, 2020 at 14:58

1 Answer 1

Reset to default
2
$\begingroup$

You could apply the GSVA transformation to later compare them. You would need to give a step back and start from the expression values not log2FC, but I think it would be better than using the log2FC directly.

If you need to use them you could compare the gene set enrichment value of these two sets on the same comparison. I would use fgsea with the log2FC and the two sets of genes. If the normalized enrichment value (NES) is different then you could explain that these sets expression are indeed different.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.