[maker-devel] Transcriptome data to gene models

[Carson Holt]

Also since single exon ESTs are almost always coming from background
transcription and not real genes, MAKER may also require some sort of
protein alignment or ab initio predictor support before promoting them to
gene models.

--Carson


From:  Sivaranjani Namasivayam <ranjani at uga.edu>
Date:  Friday, 25 January, 2013 12:17 AM
To:  Carson Holt <carsonhh at gmail.com>, "maker-devel at yandell-lab.org"
<maker-devel at yandell-lab.org>
Subject:  RE: [maker-devel] Transcriptome data to gene models

I did set the single_exon to 1, in spite of that I noticed that may single
exon transcripts don't have gene models.  I will check for other factors
like small ORFs etc, I wanted to use the est2genome just to generate quick
and gene models.  
Thanks for the information!

From: Carson Holt [carsonhh at gmail.com]
Sent: Wednesday, January 23, 2013 9:59 AM
To: Sivaranjani Namasivayam; maker-devel at yandell-lab.org
Subject: Re: [maker-devel] Transcriptome data to gene models

est2genome is really only suitable for building a training dataset for
building HMMS that will be used in a full maker run, and is not recommended
for anything else.  This is because most EST type alignments tend to be
partial.  MAKER does not try and extend these models, it just leaves them as
is, so you will end up with a lot of partial gene models.  MAKER will reject
est2genome models if they do not have splice sites (unless single_exon is
set to 1).  MAKER will also not instantiate a model if the ORF is too small
a percentage of the transcript.  Using something like apollo, you can drag
an est2genome alignment into the annotation tier and see what the model
would have looked like if instantiated.  You will probably see that for the
ones that don't produce a gene model that they either have non-canonical
splice sites or the longest open reading frame is short.

Thanks,
Carson


From: Sivaranjani Namasivayam <ranjani at uga.edu>
Date: Wednesday, 16 January, 2013 9:16 AM
To: "maker-devel at yandell-lab.org" <maker-devel at yandell-lab.org>
Subject: [maker-devel] Transcriptome data to gene models

Hi,

 I want to generate gene models for all my transcripts (based on just the
transcript evidence). To do this I set 'est2genome to gene models' attribute
to 1. I have disabled gene predictions from any another sources, that is, I
do not provide protein data and turned off all gene predictors
But I find gene models are not generated for all transcripts only about half
of them. 

Can you tell me how maker decides which transcripts to predict models for,
that is, how is it making this selective decision?

Thanks,
Ranjani
_______________________________________________ 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/20130125/607f4950/attachment-0003.html>