Gene Ontology analysis tools

[NGS_New_User]

Hi Everyone,

I desperately need your expertise advise on this

I have a assembled a genome de novo, and I would like to run gene ontology analysis on it. I have around 300k contigs. What software/tools would you advise me to try and use? I tried the free version of Blast2GO but it had limitations such that it could not work with long contigs.
Does the Blast2GO pro override that limitation?
What other GO tools would do a similar thing?

Thanks in advance

[maubp]

Normally you'd assign GO terms to genes/proteins (typically short), not entire genome contigs (typically long).

You should look at gene finding, and perhaps automated annotation in general.

[Hel]

Hi NGS_New_User,

I also have the same question. I have a de novo assembled transcriptome to analyze in gene ontology terms. Please maubp, could you explain how to get genes/proteins from our data? Because I only have contigs (and unassembled reads)..

Thanks in advance!

[rhinoceros]

Hi NGS_New_User,

I also have the same question. I have a de novo assembled transcriptome to analyze in gene ontology terms. Please maubp, could you explain how to get genes/proteins from our data? Because I only have contigs (and unassembled reads)..

Thank in advance!

Why not:

1. Predict proteins
2. Blastp against nr (tabular output)
3. Map to GO with this file
4. Sort for best hits with a GO match

[sphil]

Why not:

1. Predict proteins
2. Blastp against nr (tabular output)

or
2. Blastp against nr (XML output)
3. load into Blast2GO
4. do you analysis...

prediction of proteins you can go with augustus but there are several other programs which are capable.

[mcnelson.phd]

Hi NGS_New_User,

I also have the same question. I have a de novo assembled transcriptome to analyze in gene ontology terms. Please maubp, could you explain how to get genes/proteins from our data? Because I only have contigs (and unassembled reads)..

Thanks in advance!

You'll need to get the nucleotide sequence of all possible ORFs that are found in your contigs. To do that you can use Glimmer, GeneMark, or a few other ORF callers although the first two are probably the most popular. Then you can get a multi-fasta of all the ORFs and run those through blast2GO.

[Hel]

Thanks to all . I'm sorry but I don´t understand everything.

rhinoceros) Which tool to predict proteins?

sphill) Well, Blast2GO already do the blast step, isn't it?

mcnelson.phd) My contigs came from RNA-seq, so all my sequences have been expressed. Do you mean that I have to select the portion of the contigs that is expressed? This confuses me.

Thanks in advance!

[maubp]

sphill) Well, Blast2GO already do the blast step, isn't it?

Doing the BLAST online at the NCBI from within the Blast2GO tool is very slow. It is much faster to run the BLAST locally on a cluster, and import the BLAST results into Blast2GO for the annotation mapping step.

[sphil]

Doing the BLAST online at he NCBI from within the Blast2GO tool is very slow. It is much faster to run the BLAST locally on a cluster, and import the BLAST results into Blast2GO for the annotation mapping step.

Yep that is exactly why i should go with local blast and then blast2go.