Good one!
Hi All,
I would like to extract all the GO terms associated to a set of transcripts. I want the go association of a plant species that is not included in biomart too. I know that I can use blast2go but it takes ages for a set of 40K transcript. I have also the blast of the transcripts. is there any fast way to get all the GO terms of each transcript in a standard method?
thanks a lot :)
4 answers
To retrieve GO information from a protein dataset, I advise you to look the interproscan tool (http://www.ebi.ac.uk/interpro/search/sequence-search;jsessionid=D26B9D9396035A530BAB9546C6B0E172). I really like it.
You can try the web-service, but you can also download it and install it locally.
Once installed you just have to launch the following command:
run_interproscan --infile proteins.fa --type p --outdir interpro -appl PfamA-27.0,ProDom-2006.1 -goterms -iprlookup
Instead of PfamA-27.0 you can choose whatever domain/family/pathway/ database you want. For the GO term don't forget to use the -goterms parameter.
That really helps. Thanks @juke34.
If the genome of the species you are working with you could download the GO annotated gene from http://www.ebi.ac.uk/QuickGO/.
If the protein sequences are known for this species, and the species has GOAs, then you can link that way. Otherwise you can create your own annotations using BLASTX against a related species that has GOAs.
This is what Blast2GO does.
I think that TRAPID is one of the best solution for such a purpose. TRAPID is an online tool for the fast, reliable and user-friendly analysis of de novo transcriptomes. The GO IDs and InterPro IDs are provided for plant species.
TRAPID: an efficient online tool for the functional and comparative analysis of de novo RNA-Seq transcriptomes
Link: http://bioinformatics.psb.ugent.be/webtools/trapid/
Good Luck!
Log in to answer this question.
I also have a headsup and question. I could manage to download the Uniprotkb association to all the GO terms. Since I use uniprotkb db for my blast, is it correct if I extract the GOs associated to the proteins as the hits of my blast? and if yes, how many hits I should consider?
Is it what blast2go and other similar softwares do?