[maker-devel] maker annotate two species, same evidence, same settings

[Carson Holt]

Let me explain how ESTs and alt-ESTs are use and maybe that will answer
your question.  ESTs will be polished around splice sites to give
information on intron boundaries and UTR location.  The polished ESTs can
override ab initio gene predictor structure and be used to extend gene
models by what is basically an exon cut and paste operation.  The alt-ESTs
are not polished (exoneerate can't handle them).  So the splice site
boundaries will be off.  Maker will only try and infer UTR and intron
boundaries from these if the alignment produces canonical splice sites
(almost never happens).  So they are relegated to the same status as the
unpolished protein alignment (lower scoring hints to the gene predictor as
to exon boundaries).  They will also contribute to AED score with a little
less accuracy than ESTs.  When possible use protein models rather than
alt-ESTs, primarily because they are computationally less intense than a
tblastx type alignment that occurs with alt-ESTs.  Exonerate will also
polish protein alignments.

Thanks,
Carson



On 12-09-07 1:54 PM, "Yunfei Guo" <guoyunfei1989 at gmail.com> wrote:

>Hi Carson,
>
>I have a quick question about maker annotations. I want to annotate
>two sets of assemblies from two related species using the same sets of
>evidence and settings except that the est assemblies would be
>specified as 'est' for one species (along with 'altest' evidence from
>other species) and as 'altest' (combined with est evidence from other
>species) for the other, would the annotation reflect the real
>difference between these two species? Or there might be some
>variations within the annotations themselves which might make the real
>difference more or less significant?
>
>Thank you!
>
>Yunfei
>
>_______________________________________________
>maker-devel mailing list
>maker-devel at box290.bluehost.com
>http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org