Discrepancies between GO hits using GOATOOLS and the geneontology.org website

Question

[Yep]

I just wanted to share an issue on GitHub here and see if I can get some help from the general community. To recap what I wrote on GitHub, basically, there are some discrepancies between the results (in terms of # of ontology terms I get and also # of genes) derived from GOATOOLS and from the website. I used go.obo (downloaded from the official site) instead of go-basic.obo to avoid data discrepancies.

Please refer to the GitHub issue for details. Would appreciate any corrections/suggestions/advice/pointers!

[Jeremy]

I believe goatools and geneontology.org are two different things. geneontology.org uses PANTHER to do gene ontology analysis, so I would expect the results to be different from goatools. I've provided a link to the goatools website below.

goatools

[Yep]

Thanks for your response! They did mention on GitHub that they Process the obo-formatted file from Gene Ontology website regarding ontology. They also link the source of associations to the Gene Ontology website here, so I would be surprised if they used other sources.

In fact, the GO enrichment analysis on the Gene Ontology website is powered by Panther, as can be found here, so there shouldn't be such huge differences as I observed even if they used Panther.

I'm guessing there's probably some default settings in GOATOOLS that made it different, but I just can't figure out what that is.