Available pipelines for functional annotaiton (COG, KEGG, GO) of genomes

Question

[dago]

I am performing a comparative analysis on newly seqeunced bacterial genomes.

I annotated them using PROKKA, it worked really well.

I would like now to annotate them in respect to GO, COG, KEGG.

I was looking for a software that could do that. I Just found AutoFACT that integrates blast against different DB to get at the end a functional annotation of the seqeunces.

Has anyone experience with this software?

The installlation seems to be quite tedius because all the DB have to be downloaded locally.

Do you know an alternative pipeline?

[Martombo]

You can try blast2GO. it blasts your sequences on a huge database with annotations of GO terms, kegg pathways and interpro domains.

[dago]

Thanks. blast2GO is a good one. Do you know anything for COG?

[deni.ribicic]

I realize this was posted some while ago, but it would be nice to have your opinion on the following matter. So, I did annotation of my metagenomes using PROKKA. Is it a valid approach to use PROKKA annotated ECs and convert them to GO terms just by matching it against ec2go list?

Thanks, Deni

Answer 1

[hfang.shanghai]

If you want to use structional information (protein domains), you can try dcGO Predictor (http://supfam.org/SUPERFAMILY/cgi-bin/dcpredictormain.cgi). There are dozens of ontologies for your further consideration, which might increase your research scope.

[dago]

great thanks! I tried alread, but I submitted 100s seq and only one was annotated.

[hfang.shanghai]

Sorry for that. Now all sequences you are providing must be included in the database SUPERFAMILY.

Another way to do so is first get domain assignments from the SUPERFAMILY or similar databases. Then, use the dcGO open-source R package called dcGOR, if you are familiar with the R software environment (how to install can be found in http://supfam.org/dcGOR/install.html). You can find the detail and example at http://supfam.org/dcGOR/dcAlgoPredictMain.html.

Answer 2

[Siva]

As an alternative, you can use COGNITOR program of the COG making software COGsoft from NCBI. This basically searches the sequences (in FASTA format) from the newly sequenced bacterial genome against pre-constructed COGs and assigns the sequences to appropriate COGs.

You will also need to download the bacterial COGs to search your newly sequenced bacterial genomes against.