This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Braker Gene Annotations Post-processing

Hi,

So i have some fungal genome assemblies, for which i did

  • 1 TE annotation and assembly softmasking using EarlGrey
  • 2 Gene structural annotations using BRAKER3 un ETP mode using both RNASeq + Proteins as evidance.

My annotations are complete and i got .gff3, .gtf .aa files. and looking at basic stats i get

grep -c "gene" braker.gff3 : 16246
grep -c "mRNA" braker.gff3 : 30901
grep -c "CDS" braker.gff3:   86534

I understand the part that higher number of mRNA indicates multiple transcripts.

My next tasks are to perform Functional annotations, using tools like antismash, interproscan etc.

Questions

  • 1 Do i need to perform anz kind on QC on the annotations as there are multiple proteins in braker.aa files with X either in start or end of sequences and many protein sequences are kinda really small. for example

    g2319.t1 MDAKLTIADX

    g3697.t1 XC*

    g4306.t1 XISWACAG*

this somehow is creating issues in during the functional annotation steps.

Should i follow the "1 gene 1 protein" method and generate and filter by keeping only the longest isoforms ? if i do this then how to generate the protein sequences for the longest isoforms only ?

Thank you.

annotations braker3 fungi
• 1,625 views • link
updated by GenoMax GenoMax • written by SomeOne • SomeOne

Thank you, I got that point, but is this QC or selection of longest isoforms important step to do, Should i remove those really small proteins (the example above) from my data ?

• 0 views • link
written by SomeOne • SomeOne

What are the X symbols in your data ? I think the * symbols will cause issues downstream in a lot of applications so would remove them. I would also remove any microproteins of less than 30 or 50 amino acids. Check what others have done in your field.

Transdecoder and td2 can create the longest ORFs and protein sequences from transcript sequences if you are prepared to head further down this route.

• 0 views • link
written by colindaven colindaven

* means stop codon in some annotation tools. and what i found by googling X means ambigous amino-acid.

Check what others have done in your field

tried looking in papers which used Braker3 for annotations, but could not find anything to this point specifically.

I will try looking on Transdecoder and td2.

• 0 views • link
written by SomeOne • SomeOne

As long as you identified good gene models (based on discussion in other threads), they should be enough to allow you to create the pangenome (especially if you use a published RefSeq genome as an anchor). colindaven has more experience in this area so may have more to say.

• 0 views • link
written by GenoMax GenoMax

0 answers

No answers yet.

Log in to answer this question.

More posts like this