I am currently analysing data acquired from RNA Sequence from a specific cell type (microglia) from mouse brain. I have performed differential expression gene analysis by using DESeq2 between my control and my treatment. I would like now to perform some type of gene set enrichment analysis. According to the tutorials I can either perform fgsea or egsea.
Are there any other tools that I can use?
If I want to use fgsea:
what rank can is it better to use? I was thinking to use either the simple logFC or the signed fold change (ranked by foldFC multiply with the p values). However for the second option I don’t know how to do it without R
-Moreover, I am only working with only a cell type, will I be able to use the mouse hallmark set, or any other online available set?
I have similar question also for the egsea tool. Working only with a cell type, can I use the MSigDB gene set which is available in the tool?
For your questions → the single cell tutorials and transcriptomics tutorials have some overlap. You can certainly mix methods: QA and sorting/counting steps are distinct from higher levels of data reduction as far as I know. The complex indexes you mentions are used for both. Maybe try a few different ways and compare if you are not sure if some index is a fit for your data. The species involved, and not the sequencing methods, will matter after a certain point.
I have covered some of these tutorials from the link that you mentioned!
My main problem is when it comes to my samples and experimental design! I am really new to the analysis of the data!
I am currently working with this tutorial: 3: RNA-seq genes to pathways (3: RNA-seq genes to pathways).
And in this tutorial they are mentioning specific pre ranking options for the fgsea tool but I am not sure which one would be better for me.
Morever, I am also wondering for the gene set! both the fgsea and the egsea works with mouse set available online! However I am not sure if working with just one cell type means that I can’t work with the available data sets!
I hope that these details make my question a little bit more clear!
Maybe I am misunderstanding … I don’t think the cell type matters for this level of annotation. Known genes are expressed in all cells – or not – that’s what differential expression tools are comparing.
Ranking is about de-prioritizing genes with high expression that you may not care about, to allow for better resolution of low expressed that might be more interesting. The species and tissue (cell type) likely matter when making decisions about this.
So, publications seem like a good resource, or a scientific forum. This forum is mostly about solving technical issues. The tutorial’s cover one specific use-case and that is usually from one specific publication. For scientific opinions, and current scientific methods, you might get lucky here, but who you are asking is from a very small pool of scientists who may or may not have expertise in your domain.
The GTN hosts a chat, and that involves scientists doing analysis out in the wild. They volunteer with Galaxy … usually in their area of expertise but not always. The link is on the homepage of the tutorial site. A link to this topic can be included for context, and to provide a space for a longer reply. You’ll probably need to explain more about the data source/protocols and what your goals are. The answer will not be exact, and will probably be links to publications they happen to know about.
