[maker-devel] Masking is causing problems with exon/CDS boundary predictions

[Carson Holt]

What do your evidence alignments look like? i.e. assembled mRNA-seq and protein homology? Not the mRNA-seq pileup (because MAKER can’t see that).

Masking is applied before evidence seeding and the first ab initio run. It is actually then removed for evidence polishing around splice sites, and the hint based rerun of Augustus. So evidence is allowed to extend through masked regions and Augustus can make it part of the model if it wants on the second run. Since it didn’t the question becomes, what does the transcript and protein homology alignments from the maker run look like.

—Carson





> On Feb 8, 2017, at 12:22 AM, Salim Bougouffa <mjfi2sb3 at gmail.com> wrote:
> 
> Hi,
> 
> I am having trouble with MAKER/AUGUSTUS annotation.
> 
> One of the recurrent ones is the example I attach in artemis screenshot. In red is the gene of interest. There are three GFF files. Top is augustus standalone that I executed on the non-masked scaffold. The second is maker annotation and third is the augustus (from maker) on the masked query.
> 
> 
> The gene clearly has four exons and three CDS but the masking seem to somehow cause augustus to get the boundaries wrong for the first and third CDS and skip the second CDS entirely.
> 
> How do I fix this problem?
> 
> Best,
> /SB
> <artemis.png>
> -- 
> ____________________________
> Sent from Inbox Mobile
> 
> _______________________________________________
> maker-devel mailing list
> maker-devel at box290.bluehost.com
> http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20170208/7ce10ea3/attachment-0003.html>