Entering edit mode
Dear Bioconductors,
I have been using goseq for GO term and KEGG pathway enrichment analysis for RNAseq data. My question is if it is possible to use goseq for the same enrichment analyses for Biocarta, Reactome, and NCI pathways?
The graphite package is able to provide these pathways for topological pathway analyses, but I would like to use these pathways for the goseq enrichment approach (where gene length is taking into account).
Is there a simple way to combine or integrate these analyses (goseq with graphite)? Has anyone any experience with this combination?
Thanks in advance!
Benjamin Nota, PhD
Department of Blood Cell Research
Sanquin Research
Hello,
I am using goseq to do reactome pathway enrichment analysis. I am using the reactome.db package to obtain the link between ENTREZ gene ID and reactome pathways ID and then run the grepKEGG function as described in the goseq vignette to map ENSEMBL ID with reactome pathway ID. And then I am following the goseq tutorial for the remaining of the analysis. I am not using graphite neither Biocarta or NCI pathways but if you are able to get a database looking like to reactome.db, you should be able to map ENSEMBL ID with the pathways of your choice. However, I am not bioinformatician, so I can not guarantee this is the best way to do, I hope other people will complete my answer (I am just beginning my reactome analysis too...).
If you want to use the reactome.db package, I recommend you to manually install the reactome.db_1.48.0 version (not the last one, reactome.db_1.50.0), in other case you will have difficulties to map reactome pathways ID to reactome pathway names (see "reactome.db: reactome IDs not mapped to pathway names" topic on bioconductor support website).
Hope my answer will help you.
Best,
Nicolas
Hi Nicolas,
Thanks for your suggestion. It works! I managed to work out the approach you suggest with the reactome.db.
I guess there are no similar .db packages for Biocarta or NCI available, or are there? Or does anyone know how to convert it from graphite? Or is that impossible?
Regards,
Ben
Hi Ben,
I did not find a package containing all the biocarta and NCI database. However, if you can export from graphite or from an internet web site a file associating each pathway of the database with ENTREZ or ENSEMBL gene ID, you should be able to run goseq with such database. If it is not using ENSEMBL ID, you can use packages such as org.Mm.eg.db or biomart for conversion.
Best,
Nico