Gene Set (enrichr) and Gene Interaction Data for CEMiTool

Question

[BELA]

Hi All!

I am really new to this field, so I hope some Bioinformatics/ Biologist may help me, even if my question probably sounds silly to you.

I want to build a Coexpression network (using CEMiTool and/ or WGCNA). I think I now understood the basic concepts and I´ve already build my network. But now I want do do functional enrichment analysis and also include gene interaction data with CEMiTool.

My problem: Where do I get the gene sets in .gmt format from and where do I get the gene interaction (protein interaction) data from? At this point I think I don´t really get the difference/ purpose of all the different databases and appraoches. In WGCNA for example, they seem to use DAVID. In CEMiTool, I´ve read about GeneMania and reactome.

My data is microarray expression data from human blood cells (monocytes), 60 humans divided into 2 sample groups. With filtering I have ~5000 input transcripts for the network in CEMiTool. Would I use the Gene Symbols of these transcripts as input for the databases, to generate the tab and gmt files?

Thank you for any help!

Answer 1

[pedrostrusso]

Hi BELA, thanks for using CEMiTool!

You can get some .gmt files from the MSigDB Collections available here, including Reactome pathways. You can download several types of interaction files from a database such as STRING.

As for the format, you should make sure that both your expression file and the gmt/interaction files use the same type of identifier, or else CEMiTool will be unable to correctly identify enriched pathways, etc. If you have different identifiers, you can use a package such as biomaRt to "translate" between different identifier types.

[KOfori]

Hello Pedrostrusso I reckon you're the first author on the CEMiTool paper. Which pathways are used in the pathways.gmt file in the extdata folder of the package? Thanks

[pedrostrusso]

Hi @ko2427, yes, we used Reactome pathways, however they were edited in order to fit inside the package, so we don't recommend using them.