How to deal with missing annotations in RNASeq data?

Question

[v.berriosfarias]

Hello Community, after creating a gene transcripts per sample table from my RNASeq reads dataset from a low characterized plant species, I have performed a differential expressed genes analysis using DESeq2. The problem is I have a lot of differentally expressed genes with no functional annotations. I have used eggnogmapper to annotate the transcripts, so regarding this, What can I do in this sense? My goal is to produce a figure of the DEGs across my conditions, but if I have loosing annotations for some relevant genes I'm missing important data that I dont want to miss in the graphs. So, should I test another annotation tool? if so, which would you recommend? or should I only report those genes with annotations and exclude all those DEGs that has no functional annotation?

It will be a bad idea to discard all those DEGs that were not annotated with eggnogmapper so I will appreciate your opinions on this.

I ran eggnogmapper in the following way;

emapper.py -m diamond --itype CDS -i FASTA_FILE_NTS -o test

bests,

Valentín.

[dthorbur]

What do you mean by "low characterized plant species"? How many gene predictions does it have?

There are a few options I see:

1. If there aren't many candidates, check them with blast, convert to peptide and look at interproscan.
2. If there are many candidates, try to annotate the genome yourself and re-run analysis.
3. Use the best annotated plant genome that is similar to your target.