GO analysis: Different results EnrichGO vs ShinyGO

Question

[Clément]

Hi everyone,

I am trying to analyze my bulk RNA-seq data and perform a Gene Ontology analysis. I am wondering why my code in R gives me different results than the website ShinyGO?

Here is my code in R :

sigs <- na.omit(sigs)
genes_list <- rownames(sigs[sigs$log2FoldChange > 0.05,])
GO_res <- enrichGO(gene = genes_list , OrgDb = "org.Hs.eg.db" , keyType = "ENSEMBL" , ont = "BP", pAdjustMethod = "BH", pvalueCutoff = 0.05, qvalueCutoff = 0.2)
fit <- plot(barplot(GO_res, showCategory = 20, font.size = 8))

And I provide the gene list "genes_list" to ShinyGO: http://bioinformatics.sdstate.edu/go/

Why the results are so different and how to fix it? (see images below)

Thanks a lot

[Jean-Karim Heriche]

Does ShinyGO do the same thing with the same data as your code? There are a few different ways to deal with enrichment analysis and there are different sources of reference annotations. To start with check that ShinyGO and your code use the same version of Ensembl. Also the differences may be superficial in that the terms selected in two approaches could be related (i.e. at different levels of the same branch of the ontology), e.g. "cell-substrate adhesion" and "cell adhesion".

[Clément]

Thanks. I am using the same genes list when I run the analysis with my code or when I give it to ShinyGO, with the same ensembl names as well. The outputs have some pathways in common, but overall it's quite different, including the number of genes found related to one particular pathway.

[Jean-Karim Heriche]

That's to be expected. To clarify my previous comment, there's no reason to assume your code and ShinyGO do the same thing and use the same version of Ensembl as reference. In particular what gene is associated with which GO terms or pathways changes with different versions as the annotations are updated.