Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] BioPortal question

Trish Whetzel whetzel at stanford.edu
Fri Jun 1 11:50:49 PDT 2012


I'm done with my meeting and can dig into this further. Ketty, is 415-352-8815 the best number to reach you at?

Trish 



On Jun 1, 2012, at 11:45 AM, Mobed, Ketty wrote:

> I tried for the third time to upload the owl file (attached) to BP, with the same Bad Request error message and no upload.
> 
> Ketty
> 
> 'It's the little things that matter'
> 
> Ketty Mobed, PhD
> UCSF-IHPS-Informatics Lab
> ketty.mobed at ucsf.edu
> 415-352-8815
> From: Wynden, Rob
> Sent: Friday, June 01, 2012 10:17 AM
> To: Csongor Nyulas
> Cc: Mobed, Ketty; Natasha Noy
> Subject: Re: Fwd: BioPortal question
> 
> Okay,
> 
> We are re-testing with Csongor's method.
> 
> code and resulting OWL file are attached.
> 
> Ketty, can you please post this?
> 
> Thanks!
> Rob
> 
> 
> On 6/1/12 12:02 AM, Csongor Nyulas wrote:
>> 
>> Hi Ketty and Rob,
>> 
>> I think I discovered what the problem was. 
>> Both dbs:Table and dbs:Column (where the 'dbs' prefix stands for the 'http://www.dbs.cs.uni-duesseldorf.de/RDF/relational.owl#' namespace) are subclasses of the owl:Class metaclass, hence (1) they are also both metaclasses, and (2) their instances should be (indirect) instances of owl:Class, which means that they should show up in the class hierarchy both in Protege and BioPortal. 
>> However, apparently BioPortal shows in the class tree only direct instances of owl:Class, and not its indirect instances. 
>> 
>> I may have had an issue with the indirect subclasses of owl:Class before (I don't remember exactly), which may have prompted me to make all instances of dbs:Table also instances of owl:Class (basically double typing it), so that is why they showed up in BioPortal. It seems, however, that the same was not happening for the instances of dbs:Column, i.e. they were only instances of dbs:Column but not also of owl:Class. To fix this, I have manually added owl:Class as a second type to all the instances of dbs:Column. I used this script in the ProtegeScript plugin to set the additional types:
>> 
>> owlClass = kb.getCls("http://www.w3.org/2002/07/owl#Class");
>> colClass = kb.getCls("http://www.dbs.cs.uni-duesseldorf.de/RDF/relational.owl#Column");
>> for c in colClass.getInstances() : c.addDirectType(owlClass);
>> 
>> The result is in the attached OWL file, which you can upload to BioPortal, and you should be able to do the mappings.
>> 
>> I have to mention, though, that I am not sure how you generated this ontology, since I am not able to regenerate similar ontologies with any of the DataMaster options.
>> 
>> Csongor
>> 
>> PS: Sorry for the delayed reply, but today I was helping out at the Protege short course the entire day, and I did not have enough time to investigate this issue in depth during the day.
>> 
>> 
>> On 5/31/2012 10:57 AM, Mobed, Ketty wrote:
>>> 
>>> I have updated the permissions. Here is the ontology link
>>> 
>>> http://bioportal.bioontology.org/ontologies/47565 
>>> 
>>> Thanks,
>>> Ketty
>>> 
>>> From: Csongor Nyulas [csongor.nyulas at stanford.edu]
>>> Sent: Thursday, May 31, 2012 10:52 AM
>>> To: Mobed, Ketty
>>> Cc: Wynden, Rob; Natasha Noy
>>> Subject: Re: Fwd: BioPortal question
>>> 
>>> It is: csnyulas
>>> Thanks,
>>> Csongor
>>> 
>>> On 5/31/2012 10:48 AM, Mobed, Ketty wrote:
>>>> 
>>>> Hi Csongor,
>>>> 
>>>> can you give me your BioPortal username and I will add you to get access to this ontology.
>>>> 
>>>> Thanks,
>>>> Ketty
>>>> 
>>>> 'It's the little things that matter'
>>>> 
>>>> Ketty Mobed, PhD
>>>> UCSF-IHPS-Informatics Lab
>>>> ketty.mobed at ucsf.edu
>>>> 415-352-8815
>>>> From: Csongor Nyulas [csongor.nyulas at stanford.edu]
>>>> Sent: Thursday, May 31, 2012 10:06 AM
>>>> To: Wynden, Rob
>>>> Cc: Natasha Noy; Mobed, Ketty
>>>> Subject: Re: Fwd: BioPortal question
>>>> 
>>>> Hi Rob,
>>>> 
>>>> On 5/31/2012 12:03 AM, Wynden, Rob wrote:
>>>>> 
>>>>> Hi,
>>>>> 
>>>>> The database is too large to send and comprised of very sensitive patient data.
>>>> 
>>>> I was thinking of a small sample, but I suppose that even just sending the structure of the tables (i.e. empty tables, or SQL statements that create the empty tables) would be sufficient. Only in the case that you think that the DataMaster plug-in does not work properly, and you would like me to debug it.
>>>> 
>>>>> 
>>>>> The problem is that if the Owl file is uploaded to BioPortal then we cannot use Term Mapping on the column names.  The columns are classes but for some reason the Term Mapping tab doesn't work.
>>>> 
>>>> I think I understand. Can you send me a link to the ontology uploaded to BioPortal, to take a look at it? Is it a public ontology?
>>>> 
>>>> Csongor
>>>> 
>>> 
>> 
> <UCSF_bpds-vdb_schema_1489431912481927878.owl>

Trish Whetzel, PhD
Outreach Coordinator
The National Center for Biomedical Ontology
Ph: 650-721-2378
http://www.bioontology.org

"Like" NCBO on Facebook: http://on.fb.me/bioontology

Follow NCBO on Twitter: http://twitter.com/#!/bioontology

Join in Discussions on LinkedIn: http://linkd.in/ncbo-group

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20120601/f60a848c/attachment.html>


More information about the bioontology-support mailing list