[maker-devel] gff3 valiadation with GAG and loss of Name information, start-stop codons

[Carson Holt]

Final gene models will have ‘maker' in the source column. All others are not gene models, but rather evidence alignments that are there for reference. Make sure you specified a gene predictor during the maker run, or you may only have alignments and no models. Anything with protein_gff:protein2genome in the source column would fall into the category of evidence alignment.

I do not believe GAG supports tRNA’s. Someone may be able to correct me, but I believe they only support the coding annotations.

—Carson




> On Feb 23, 2018, at 7:43 AM, Giroux, Emily (CFIA/ACIA) <Emily.Giroux at inspection.gc.ca> wrote:
> 
> Hi,
>  
> After running my rounds with Maker, I went on to using the Genome Annotation Generator (GAG) and tbl2asn to find errors and fix them. My plan is to run BLAST and InterProScan after so that I don’t waste time annotating invalid features. However, when running GAG, all the protein_gff:protein2genome annotations get removed, along with the Name matches that I got to related species. Somehow I feel it would be better to find the errors, and go back to the last Maker generated gff3 instead of continuing on successively with each GAG-cleaned-up gff. However, another thing is that the start and stop codons are not included in the Maker gff, and there is no way that I can fix this all in the original manually.
>  
> My question is, is it a real loss to lose the protein_gff:protein2genome information, and all the Name info, or does this get added back once I run BLAST and InterProScan? Does Maker have a function to add the Start and Stop codons? Should I really be running BLAST and InterProScan before going on to GAG? Lastly, MAKER adds tRNA annotations, but for each of these, tbl2asn results in errors: SEQ_FEAT.MissingTrnaAA, and I’m not sure if I can just ignore this…
>  
> Thanks,
>  
> Emily

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20180226/645ffb5f/attachment-0003.html>