[maker-devel] Question about AED scores

[Carson Holt]

AED from MAKER in the based on nucleotide-level congruency of the transcript
compared to the collapsed evidence.

eAED is also at the nucleotide level but is adjusted for inferred support of
exons, such as can happen with mRNAseq (I.e. I can confirm the splice sites
at each end but as the exon gets longer the middle is less likely to be
covered).  So if I can infer the middle based on ORF and the splice site are
confirmed then I consider the middle of the exon to be confirmed for
calculating eAED.  eAED also adjusts for reading frame from protein evidence
alignments, i.e. protein overlap in a different reading frame than the final
gene model will not contribute to eAED.  So because of this, eAED takes much
longer to calculate but is sometimes more useful.  Much of the time AED and
eAED are identical.

Thanks,
Carson


From:  Daniel Standage <daniel.standage at gmail.com>
Date:  Wednesday, 23 January, 2013 10:31 AM
To:  Maker Mailing List <maker-devel at yandell-lab.org>
Subject:  [maker-devel] Question about AED scores

Good morning!

I have a quick question or two about the AED scores reported for each
transcript annotated by Maker. So, if I understand correctly, AED is 1-C,
where the congruency C is the average of the sensitivity and specificity.
The sn and sp values can be calculated for different levels (whole
transcripts, exons, or individual nucleotides), but the text of the Maker2
paper suggests the reported AED value is calculated based on
nucleotide-level congruency with the aligned evidence--is this correct?

I'm assuming this score is stored in the _AED attribute, which leads to my
other question: what does the value of the _eAED attribute represent? Is
this the same score calculated at the exon level?

Many thanks!

--
Daniel S. Standage
Ph.D. Candidate
Bioinformatics and Computational Biology Program
Department of Genetics, Development, and Cell Biology
Iowa State University
_______________________________________________ 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/20130124/161950b9/attachment-0003.html>