<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
</head>
<body>
<div class="BodyFragment"><font size="2"><span style="font-size:10pt;">
<div class="PlainText">Ok, do you by that imply that the predictions that end up in the gff3 output from the ab initio predictors (snap_masked, augustus_masked, and genemark), are not the final hinted predictions? Otherwise, I’m sorry that I can’t follow your
reasoning. I checked my gff file, and there is no evidence there to support the bridge, as far as I can tell (See attached gff of the region or
<a href="http://figshare.com/articles/Maker_prediction/942301">http://figshare.com/articles/Maker_prediction/942301</a> where all evidence is plotted).<br>
<br>
Mikael<br>
<br>
</div>
</span></font></div>
<div class="BodyFragment"><font size="2"><span style="font-size:10pt;">
<div class="PlainText"><br>
<br>
23 feb 2014 kl. 21:00 skrev Carson Holt <carsonhh@gmail.com>:<br>
<br>
> The image doesn’t show all evidence sources, but the short answer is that<br>
> one of you evidence sources (est2genome, protein2genome, or blastx)<br>
> bridges the two regions, and when provided the bridged hint one of the<br>
> gene predictors thinks it makes sense to create a single model instead.<br>
> my guess is that it’s blastx evidence.<br>
> <br>
> —Carson<br>
> <br>
> <br>
> On 2/23/14, 9:57 AM, "Mikael Brandström Durling" <mikael.durling@slu.se><br>
> wrote:<br>
> <br>
>> Dear list and maker developers,<br>
>> <br>
>> I was browsing the results of a recent maker run, focusing on differences<br>
>> between this run with the a recent maker (svn r1067) and a previous run<br>
>> with svn revision 1022 (I recall). One of the differences I found was a<br>
>> gene lost in the new prediction set, but replaced by an extended version<br>
>> of a previous neighbor (see<br>
>> <a href="http://figshare.com/articles/Maker_prediction_comparison/942300">http://figshare.com/articles/Maker_prediction_comparison/942300</a>). As you<br>
>> can see, there is no support for the join in the evidence. Do you have<br>
>> any clue to what might cause this?<br>
>> <br>
>> Best regards,<br>
>> Mikael Durling<br>
>> <br>
>> <br>
>> <br>
>> _______________________________________________<br>
>> maker-devel mailing list<br>
>> maker-devel@box290.bluehost.com<br>
>> <a href="http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org">
http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org</a><br>
> <br>
> <br>
<br>
</div>
</span></font></div>
</body>
</html>