[maker-devel] maker output

[Carson Holt]

For AED in general, lower is better, but you have to understand the caveats.
With mRNA-seq Nnot all genes may be expressed, not all exons may be captured
(mRNA can fold blocking some sequencing reactions), and sometimes the
alignment may extend improperly into the intron or even merge into the
neighboring gene.  Also mRNA-seq captures a lot of things that aren't coding
genes.  But in general for mRNA-seq, as coverage increases the AED values
trend toward 0, and mRNA-seq is the single most informative piece of
evidence you can get for annotation (I've seen several very poor genome
assemblies with horrible annotations that were saved by mRNA-seq).  For
mRNa-seq, give MAKER the assembled reads (trinity works well).  Also for
fungi, the UTR tend to overlap between genes.  This can create false merging
in the mRNA-seq assemblies (their AED is lower but its a false merge). Use
the correct_est_fusion option in the control files to help handle that.

I know there are also several members of the MAKER mailing list who have
extensive experience using mRNA-seq to annotate fungi who may want to add
their two cents.

Thanks,
Carson





From:  Hud Hud <hudarul at yahoo.com>
Reply-To:  Hud Hud <hudarul at yahoo.com>
Date:  Monday, 15 April, 2013 11:54 PM
To:  Carson Holt <carsonhh at gmail.com>
Subject:  maker output

Hi Carson, have a nice day..
I have a question about the output file from maker, recently i run my
longest contigs (100kb) on Maker using rna-seq data from other related
species of my genome (same genus)..and ive noticed that i managed to get
expressed sequences match annotation compared using just EST and cDNA. Is
this due to different size of dataset? as im using larger dataset when im
incorporating rna-seq data (assembled transcript combined with cDNA and est)
. The value of AED for both prdicted mRNA 0.15 (with rna-seq data) and
0.06(w/o rna-seq data). My question is which one is the most accurate
prediction, can i just depends on the value of AED ( the lower the better)?
How about the incorporation of rna-seq data in this case,can i conclude that
rna-seq improves the annotation (based on the image i attached). Thanks for
your time, really appreciate it.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://yandell-lab.org/pipermail/maker-devel_yandell-lab.org/attachments/20130416/0900e8d3/attachment-0002.html>