[maker-devel] gff output

[Carson Holt]

If you are using the gff3 directly produced by Augustus, it will be oddly structured and does not conform to the 'Canonical Gene’ example given by the GFF3 format specification.  You have to make a couple of search and replace operations to make it work.

Also it would generally be better to let maker run augustus for you rather than providing it as GFF3. This is because you lose the hint feedback that maker provides augustus.  AS a result there will be no improvement made to the annotations beyond what augustus has already produced.

—Carson




> On Nov 27, 2014, at 2:22 AM, Muriel Gros-Balthazard <muriel.grosb at gmail.com> wrote:
> 
> Hello,
> 
> I have been using Maker to generate an annotation.
> I especially set these options:
> - est_gff with a list of transcripts.gff3 (Cufflinks output)
> - model_org=all
> - rmlib=allrepeats.lib
> - repeat_protein=te_prot.fasta
> - pred_gff= Augustus.gff3 (that I generated previously)
> 
> I obtain a gff file for each of my contigs.
> However, here are the three possibilities in the second column :
> # est_gff:cufflinks
> # repeatmasker
> # repeatrunner
> 
> I have no information about exons and introns.
> And I am wondering if the Augustus.gff3 was used...
> 
> On top of that, I forgot to set up pred_stats to 1.
> If I understand well, I can just change this in the ocntrol file, and run Maker again. Since there is the output with everything, it won't run again the prediction, only this option. Is that right ?
> 
> Thank you,
> 
> Muriel
> 
> 
> _______________________________________________
> maker-devel mailing list
> maker-devel at box290.bluehost.com
> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org