[maker-devel] annotation comparison aed plots

[Carson Holt]

I don’t know about Michaels’s script, but I’ve always used eval.  It
produces sensitivity/specificity metrics.  It assumes the first models are
100% correct, and then tells you the sensitivity/specificity value for the
second models.

It is not therefor a quality metric.  Instead you should view it as a change
metric. Lower sensitivity tells you that models/exons have been lost between
versions, and lower specificity tells you models/exons have been gained.
There will also be a lost of generic statistics on exon/intron distribution
and UTR length.  Then the AED values from the MAEKR run can be used
independently to evaluate how well models match the evidence.

—Carson


From:  "Robert King (RRes-Roth)" <robert.king at rothamsted.ac.uk>
Date:  Monday, March 10, 2014 at 5:17 AM
To:  "maker-devel at yandell-lab.org" <maker-devel at yandell-lab.org>
Subject:  [maker-devel] annotation comparison aed plots

Dear Maker Developers,
 
I’ve updated a reference that was had errors and was a little incomplete and
now trying to produce a annotation for it. Please note the reference has not
changed dramatically. I’ve produced two annotations using as evidence:
 
Annotation 1:
Uniprot proteins search using species keyword “fusarium”
Pubmed mRNA for the name of the organism
Prior annotation reference transcripts
 
Annotation 2:
Uniprot proteins search using species keyword “fusarium”
Pubmed mRNA for the name of the organism
Prior annotation reference transcripts
mRNA trinity assembly pasafly of different strain (only RNA-seq available)
 
I’m not sure if it was a smart move to use the prior annotation reference
transcripts?
 
I want to compare these two annotations and have produced AED scores. How do
I generate summary stats/figures to compare annotations. You mentioned last
year in a post Mike Campbell has a script to produce these, do you know if
he will post it? I’ve got the Eval program and converted to gtf format using
the provided script, just waiting on some perl modules to be installed by
admin to test it. I’m waiting on some perl modules to be installed by our
administrator to test out the “Evaluator” and “compare” programs too, what
do they do?
 
Best Wishes
Rob

-- 
This message has been scanned for viruses and
dangerous content by MailScanner <http://www.mailscanner.info/> , and
we believe  but do not warrant that this e-mail and any attachments thereto
do not contain any viruses. However, you are fully responsible for
performing any virus scanning.
_______________________________________________ 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/20140310/cbd8263c/attachment-0003.html>