How to do gene functional annotation ?

Question

[vin.darb]

Hello

As part of my bioinformatic internship, one of my objectif is to perform functional annotation on single cell sequencing data, but I'm a little lost ..

Till now, I have assembly my reads (from differents SAGs) into contigs with SPADes (after doing quality control), I have use Prodigal to predict ORF on these contigs and I have BLAST these ORF against Eggnog database.

In the tabulated output file generate by BLAST, my idea was to collect accession name from Eggnog, for exemple:

733_F2_>NODE_3161_length_96_cov_1_ID_6321_1 13616.ENSMODP00000006300 100.0 31 0 0 1 31 54 84 4.1e-09 65.1

and for each of these accessions name retrieve the biological process, molecular function etc to do some graphs using R by couting the number of occurrences in each category

But I can't find file to make the link between the accessions name and the gene ontology ..

My idea is maybe bad ? I'm a newbie in bioinformatics and I don't know how to process functional annotation, maybe I'm not at all on the right track

What would you do in my position?

Thank's in advance and sorry for my english

[Gio12]

You may consider using PANTHER or DAVID in order to find your GO terms for your accession names.

[vin.darb]

I will try, thank you

It's only via web interface or it can use on commande-line ?

[lieven.sterck]

you can also consider using tools like : blast2GO or interproscan

something like Trapid can also be helpful I think

a very crude approach would to to simply blast your sequences against uniprot/swissprot and 'transfer' the description of the best hit

Answer 1

[ddeemer]

PANTHER and DAVID both provide good resources, but I found some of their pathways to be convoluted when analyzing many genes. Also, check out the Reactome Pathway Database, I found that pretty user friendly and neat to play with.