This is a test version of Biostars. For the public version, visit https://www.biostars.org.
How to perform GO enrichment in non-model organisms?

Hello

I am a beginner using RNA-seq data for analysis. I would like to use GO enrichment analysis to infer the functions of genes that are upregulated by DEG analysis. However, since my analysis target is a non-model plant, there are no annotation files available, so I have to create them myself. Fortunately, I obtained the whole genome sequence and the GFF file describing the positions of the gene sequences. If I were to annotate all genes myself, what tools would be optimal? I personally plan to use eggNOG-mapper.

go
• 2,381 views • link
updated by bioadam • bioadam • written by CaERu _no_ korokoro • CaERu _no_ korokoro

1 answer

In first instance you will as you say need to assign GOterms to your genes. eggNOG-mapper is certainly a valid approach to do this indeed.

Other options are Interpro2GO (first run an interpro analysis, then 'convert' the results to GO-terms) or even Blast2GO . All approach have their strengths and weaknesses so it is a bit of a trade-off between precision and recall for all of them. You might consider running all approaches and combine the results afterwards to get a more complete view. Keep in mind that these can be quite heavy analyses (eg. Interproscan running) so you ideally have access to serious compute resources (and even then they might run for a while).

In any you will in the end get more useful info than only the GO-terms, Interpro will give you info on the protein domains present which to itself can also indicate function (even when lacking GO-terms). Just to say that these are valuable analyses to do anyway.

You should also have a look at GOslim: this is a kind of higher level GO annotation and might prove more useful for DEG analysis

• 0 views • link
written by lieven.sterck lieven.sterck

Thank you for your reply. Recently, I've been unable to access the eggNOG-mapper server, so I'm currently using a tool called Galaxy for analysis. It's a bit of a hassle... Indeed, when annotating with eggNOG-mapper, it was not possible to annotate all gene sequences. As you mentioned, I will try various approaches and make a comprehensive judgment based on them.

• 0 views • link
written by CaERu _no_ korokoro • CaERu _no_ korokoro

Ah, yes, Galaxy is not a bad choice here. It's a platform to easily run several tools without the hassle to get them installed/running locally. Many (all?) of the analyses I mentioned are likely available on it.

Good luck with it.

• 0 views • link
written by lieven.sterck lieven.sterck

I'm a little late to this post, but I'll add that it's completely normal, even expected, for a decent proportion of sequences to receive no annotation. This is especially true when working with non-model taxa. The exact species you’re working with, the version of eggNOG you’re using, the quality of your assembly, and even sequencing depth can impact this. I wouldn't worry too much about that in itself.

• 0 views • link
written by bioadam • bioadam

Log in to answer this question.

More posts like this