Entering edit mode
5.5 years ago
t-jim
▴
30
Hello,
It´s my first time doing RNA-Seq. I performed differential expression analysis using DESeq2 and now I want to do gene ontology analysis with my data to be able to see the differences between the two groups (control and treatment).
I tried it using goseq but this gives me only one output for everything combined. What I want is to perform a GO analysis for my control group and for my treatment group so I can compare them with each other. Does anyone have an idea?
I would appreciate any help.
Thank you for your answer! Is there a way to see which GO term is associated with which group?
You can split the up and down genes up in two different lists if you want. Genes up in treatment give you info about what processes are more induced concerning the treatment.
Sorry for the late reply. From my understanding this helps me to see which terms are up and which are down regulated, but there is no information about which terms are associated with which group.
But in your question you say:
So you want to know which GO terms are different between your groups, right?
Yes that is what I want. I want to have a list of GO terms for my treatment group and one for my control group which I both used for differential gene expression analysis. Correct me if I´m wrong but as I said your suggestion gives me a list of GO terms for all up reguated genes and one for all down regulated genes. But I don´t know which genes are up/down regulated in which group so that doesn´t solve my problem.
The definition of up regulated genes is usually; higher expression in the treatment group compared to control. Down regulated genes are higher in control. Generating a list for each group (as you suggest) is not really the way to go (no pun intended). Please read some literature where they use goseq for treatment/control experiments to see how you can use it.
Thank you again for your answer! I didn´t know that you don´t generate a list for each group (I´m new to RNA-Seq as I said and that is what I was asked to do) also I didn´t know that you can infer the group from up/down regulation. Just for curiosity, how would you do that in a multifactor experiment (e.g. 4 different groups)?
Since goseq needs all measured genes as input, I was wondering what would you use as the measured genes when performing goseq only for up regulated genes. Would you use all genes with a padj < 0.05 or only the up regulated genes with a padj < 0.05?
I always use padj < 0.05. There are many ways to use GO analysis, if you have e.g., 4 groups it depends on what these groups are (3 treatments and one control? Or 4 cell populations?), and what your research question is. Usually if you want to do GO analysis on more groups, you can use clusterProfiler.