Question: How to make sense of the fold change values in the interaction results object.
1
3.2 years ago by
colaneri30
United States
colaneri30 wrote:

Hi

I have and experiment with 2 genotype and 2 conditions. i was interested in identify genes with genotype:condition interactions. The res objet report p-values but also log fold change. How can I make sense of these fold changes?  At least how can interpretive negative from positives log fc?

I will appreciate any insight

ALe

modified 3.2 years ago by Michael Love23k • written 3.2 years ago by colaneri30
Answer: How to make sense of the fold change values in the interaction results object.
0
3.2 years ago by
Michael Love23k
United States
Michael Love23k wrote:

You should first read the section of the vignette on how to interpret interaction terms.

vignette("DESeq2")
1

Dear Michael, I have  read the vignette, but there is not explanation on how to interpret fold change values in a model that tested interactions. Given that interactions measures significant difference in fold change for a same gene across different conditions (two in mi case) I do not understand what does the reported fold change means. It can not be an average of the compared fold changes neither one or the other. So here my question again. How can I make sense of it?

ALe

I'm not sure I 100% understand the question.

There is a fold change associated with the reference level.

It would be easier for me to answer you if you could refer to specific terms that you do not understand, say using the example depicted in Figure 9 in the Interactions section in the vignette, or using the examples of interaction terms that are described in the Examples section of ?results.

Hi Michael

I think now that I read vignette 2016 I get the answer.

Thank you!

Make sure you've updated to the latest version of DESeq2! It's important, as Bioconductor developers spend a lot of time improving software and fixing bugs. You should update before starting any new analyses. This involves installing the latest version of R from CRAN and then following these instructions:

http://bioconductor.org/install/

Also, note that a new version of Bioconductor will be available in 3 weeks, tied to R 3.3 (not yet released). Bioconductor is updated to a new version every 6 months.