[ensembl-dev] Attaching DAS, species not found error

Andy Jenkinson andy.jenkinson at ebi.ac.uk
Wed Nov 10 15:26:31 GMT 2010


Hi Ken,

I'm afraid without doing some debugging at your end it's very difficult to figure out what's going on, as I can't reproduce the problem. You could try turning on logging at the INFO level, and adding some debug statements around the "fetch_Sources" call in ensembl-webcode/modules/EnsEMBL/Web/Object/UserData.pm to figure out what filters the web code is sending to the DAS parser to explain why no sources are returned. Likewise in the parser itself (ensembl-external/modules/Bio::EnsEMBL/ExternalData/DAS/SourceParser.pm). If you contact me off-list with your findings I will help you get to the bottom of it.

As an aside, the trinomial species issue is likely to be solved soon as Ensembl Bacteria have a similar problem. The plan is to match using the taxonomy ID instead of the species name. It's important for compatibility throughout DAS (i.e. not just Ensembl) to use the names from the NCBI taxonomy in DAS servers and handle them in clients, so it will be great to get this fixed so you don't have to hack it!

Cheers,
Andy

On 9 Nov 2010, at 20:09, Ken Youens-Clark wrote:

> Andy,
> 
> You're right, my coordinates were messed up for several species.  They're not all perfect, but I hacked my registry_coordinates.xml file to change "MSU_6,Chromosome,Oryza sativa japonica" to "MSU_6,Chromosome,Oryza sativa" since our japonica browser has to be called "Oryza_sativa."  The first is actually more correct, but there really seems to be a problem with DAS handling the trinomial names -- it's also a problem for our "Oryza sativa indica" browser which we have to call just "Oryza_indica."  
> 
> So now this looks like it /should/ work:
> 
> 	http://www.gramene.org/gramenedas/das/sources
> 
> But I still get the same error with "Oryza_sativa" not being found.  Out of curiousity, I also added "organism = Oryza sativa" to one of my sources, but that didn't help, either.
> 
> Gramene is using Ensembl v59 right now.
> 
> Any suggestions?
> 
> ky
> 
> On Nov 9, 2010, at 10:59 AM, Andy Jenkinson wrote:
> 
>> Hi Ken,
>> 
>> At first glance: your DAS server does not report any coordinate systems for the Oryza (and others) sources, like it does for some other species'. Ensembl therefore doesn't know that they are for Oryza. If you need to add suitable coordinate systems to the registry, an admin of the DAS Registry can help (me included).
>> 
>> But actually, the behaviour you describe does not reflect what happens on the main Ensembl site. If the source has no coordinate systems, here they are still listed but simply ask you to specify the correct coordinate system. This seems to work, so the problem may be specific to gramene. Take a look at:
>> ensembl-webcode/modules/EnsEMBL/Web/Object/UserData.pm
>> ensembl-external/modules/Bio::EnsEMBL/ExternalData/DAS/SourceParser.pm
>> 
>> Which version of Ensembl is gramene based on?
>> 
>> Cheers,
>> Andy
>> 
>> On 9 Nov 2010, at 16:19, Ken Youens-Clark wrote:
>> 
>>>> From Gramene's rice browser (Oryza sativa japonica):
>>> 
>>> 	http://dev.gramene.org/Oryza_sativa/Location/View?r=1:8001-18000
>>> 
>>> I try to attach a DAS track via "Manage your data" -> "Attach DAS" and specifying "http://dev.gramene.org/gramenedas/das" as the URL, I get this error:
>>> 
>>> 	Unable to access DAS source. Server response: No Oryza_sativa DAS sources found for http://dev.gramene.org/gramenedas/das
>>> 
>>> But if I look here:
>>> 
>>> 	http://dev.gramene.org/gramenedas/das/sources
>>> 
>>> I do see "Oryza_sativa" sources.  How or where exactly is the browser looking for "Oryza_sativa" to know what it can display?
>>> 
>>> --
>>> Ken Youens-Clark
>>> kclark at cshl.edu
>>> Gramene.org Project Manager
>>> Ware Lab/CSHL/USDA-ARS
>>> 
>>> 
>>> _______________________________________________
>>> Dev mailing list
>>> Dev at ensembl.org
>>> http://lists.ensembl.org/mailman/listinfo/dev
>> 
> 
> --
> Ken Youens-Clark
> kclark at cshl.edu
> Gramene.org Project Manager
> Ware Lab/CSHL/USDA-ARS
> 
> 
> _______________________________________________
> Dev mailing list
> Dev at ensembl.org
> http://lists.ensembl.org/mailman/listinfo/dev





More information about the Dev mailing list