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] A few questions on Bioportal Sparql Endpoint

manuelso manuelso at stanford.edu
Thu Mar 7 10:06:33 PST 2013


Hi Hong Sun.


> Thank you very much for your detailed explanations! I now have a much better understanding, and yet still a few questions would like to check with you concerning the matching. 

You are welcome. Find below the answers inlined.

> 
> First, in my last email, I mentioned that Is it possible to add skos:inScheme for instances like below? 
> You can open the URL I enclosed, the query I executed is : 
> Query1: 
> PREFIX maps: <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#> 
> SELECT * 
> FROM <http://bioportal.bioontology.org/ontologies/ICD10> 
> WHERE { 
>    <http://purl.bioontology.org/ontology/ICD10/I21> ?p ?o } 
> I am wondering it would be nice to have a skos:inScheme property for this instance, is it possible to add? 

I will bring this question to the rest of the team. I am not sure if we can add a triple per term stating the scheme they belong to.

> 
> Second, a few questions concerning mapping. 
> 1. 
> Your answer below well explained my question, 
> "Here there is a table with the types of mappings we host (http://www.bioontology.org/wiki/index.php/SPARQL_BioPortal#mappings) .  REST mappings are submitted by users, CUI mappings are released with the UMLS database and are human-curated, LOOM are lexical mappings (terms with equals labels/synonyms), ..." 
> I get the explanation from the web page: 
> 
> <Mail Attachment.gif> 
> I found the graph name is implicitly stated in each mapping instance, 
> e.g. a mapping instance <http://purl.bioontology.org/mapping/umls_cui/9edef869c7003a554b6e0a59547c451b> tells it is from umls_cui graph. 
> Is there any explicit way to retrieve the mapping type in the above table? 
> Is it possible to add a property in the mapping ontology to directly/explicitly state these different mapping types? 

You are right. At the moment we do not have that predicate, the type of the mapping is conceptualized by the graph that contains it. As part of a separate effort, a new version of REST APIs, we are remodeling the mapping data. We will take this into consideration. Good point.

> 
> 2. I queried the following query at the endpoint, 
> Query2: 
> PREFIX map: <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#> 
> SELECT * WHERE { 
>       ?m map:source <http://purl.bioontology.org/ontology/ICD10/I21> ; 
>              map:target  <http://purl.bioontology.org/ontology/SNOMEDCT/194796000> . 
>   ?m ?p ?o. 
> } 
> 
> I get a triple 
> <http://purl.bioontology.org/mapping/umls_cui/9edef869c7003a554b6e0a59547c451b>         <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#dependency>         <http://purl.bioontology.org/mapping/umls_cui/7f5ffe5eb1e0db2536db9bf810ac6c55> 
> What does <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#dependency> mean? 

One mapping "depends" on another mapping when they connect the same terms but with different directions. Basically the target and source are shifted.

> 
> 3. I found that in the published mapping ontology http://protege.stanford.edu/ontologies/mappings/mappings.rdfs# 
> there are many interesting properties, like mappings:confidence mappings:authority mappings:verified mappings:submittedBy mappings:verified_by 
> query these properties in the endpoint does not return any result yet, would they be supported in the later versions? 

That ontology defines relationships for mappings in a broader context. At the moment we do not have any actions planned to generate data using those predicates.

> 
> 4. I found there are some inconsistence with the ontologies among the 
> published ontology, http://protege.stanford.edu/ontologies/mappings/mappings.rdfs# 
> introduction webpage,  http://www.bioontology.org/wiki/index.php/SPARQL_BioPortal#mappings 
> and the endpoint. 
> 
> e.g. 
> in the introduction web page, there is a mapping instance 
> @prefix maps: <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#> .
> 
> <http://purl.bioontology.org/mapping/2767e8e0-001b-012e-749f-005056bd0010>
>    maps:has_process_info <.../procinfo/2008-04-23-38138> ;
>    maps:comment "Manual mappings between Mouse anatomy and NCIT." ;
>    maps:relation skos:closeMatch ;
>    maps:target <http://purl.org/obo/owl/MA#MA_0001096> ;
>    maps:source <http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl#Olfactory_Nerve> ;
>    maps:source_ontology_id <http://bioportal.bioontology.org/ontologies/1032> ;
>    maps:target_ontology_id <http://bioportal.bioontology.org/ontologies/1000> ;
>    a maps:One_To_One_Mapping . 
> 
> It uses maps:source_ontology_id 

right. We need to update the published ontology. Thanks for pointing this out.

> 
> In the endpoint, it uses mappings:target_ontology 
> e.g. below is a triple returned from Query2 
> <http://purl.bioontology.org/mapping/loom/9edef869c7003a554b6e0a59547c451b>         <http://protege.stanford.edu/ontologies/mappings/mappings.rdfs#target_ontology>         <http://bioportal.bioontology.org/ontologies/1353> 
> 
> And in the published ontology http://protege.stanford.edu/ontologies/mappings/mappings.rdfs# 
> neither maps:source_ontology_id nor mappings:target_ontology  is mentioned as a property. 
> 
> Again, thank you very much for your kind help! 

Thanks to you. Vey good questions indeed.

Manuel

> 
> Kind Regards,
> 
> Hong Sun | Agfa HealthCare
> Researcher | HE/Advanced Clinical Applications Research
> T  +32 3444 8108
> 
> http://www.agfahealthcare.com
> http://blog.agfahealthcare.com
> Click on link to read important disclaimer: http://www.agfahealthcare.com/maildisclaimer 
> 
> 
> 
> From:        manuelso <manuelso at stanford.edu> 
> To:        Hong Sun/AXIFX/AGFA at AGFA 
> Cc:        "support at bioontology.org Support" <support at bioontology.org> 
> Date:        02/20/2013 07:26 PM 
> Subject:        Re: A few questions on Bioportal Sparql Endpoint 
> 
> 
> 
> (Hong Sun kindly agreed to have this response sent to the support list) 
> 
> Hi Hong, 
> 
> See inlined answers below ... 
> 
> On Feb 18, 2013, at 8:53 AM, Hong Sun <hong.sun at agfa.com> wrote: 
> 
> This is Hong Sun, I am a researcher working on AGFA healthcare researching the semantic web in clinical domain. 
> I have read your paper: BioPortal as a Dataset of Linked Biomedical Ontologies and Terminologies in RDF, 
> and practiced your BioPortal SPARQL endpoint. It is great to have such a place where you can query different terminology and the mappings among them. 
> 
> Thanks! 
> 
> 
> I have three questions would like to check with you: 
> 1. As most properties in the endpoint uses SKOS properties, why do you chose the rdfs:subClassOf, instead of skos:broader, to represent the hierarchy? 
> 
> The ontologies that BioPortal hosts have been mostly developed by other institutions.  It is the author's choice to use rdfs:subClassOf to represent the hierarchy. At the moment BioPortal only supports rdfs:subClassOf and that is why you probably will not see that many SKOS relations in the SPARQL endpoint. We are going to support SKOS very soon and that will probably change. 
> 
> 2. As the typed literals for skos:notation are stated as "best practice" of SKOS, is it possible to use typed literal to represent notation? 
> 
> We generate some skos:notation predicates but not all of them. The ones we generate are typed, we do not have control over the literals generated by others. 
> 
> Our Web frontend (sparql.bioontology.org) hides the literal types, so probably you cannot see them there. If you query the endpoint programmatically you'll be able to see the literal typed. 
> 
> Some example here: https://github.com/ncbo/sparql-code-examples 
> 
> 
> 3. As the mappings between different terminologies are submitted by different people, to which extent can I trust the mapping? 
> 
> Every mapping has its process/provenance information. You can look at that metadata and decide how much you can trust a single mapping or a collection of mappings--it is up to you or the application. Here there is a table with the types of mappings we host (http://www.bioontology.org/wiki/index.php/SPARQL_BioPortal#mappings) .  REST mappings are submitted by users, CUI mappings are released with the UMLS database and are human-curated, LOOM are lexical mappings (terms with equals labels/synonyms), ... 
> 
> 
> 4. Is it possible to add skos:inScheme for instances like below? 
> http://sparql.bioontology.org/?query=PREFIX+maps%3A+%3Chttp%3A%2F%2Fprotege.stanford.edu%2Fontologies%2Fmappings%2Fmappings.rdfs%23%3E%0D%0ASELECT+*+%0D%0AFROM+%3Chttp%3A%2F%2Fbioportal.bioontology.org%2Fontologies%2FICD10%3E%0D%0AWHERE+%7B%0D%0A+++%3Chttp%3A%2F%2Fpurl.bioontology.org%2Fontology%2FICD10%2FI21%3E+%3Fp+%3Fo+%7D&kboption=ontologies&csrfmiddlewaretoken=3d749d07c0ac5d0a0bff52fdcae2c04e 
> 
> 
> Sorry I cannot see the query there … can you copy & paste the query in an email. 
> 
> Good questions! and thanks for the feedback. 
> 
> Manuel 
> 
> 
> Thank you very much in advance! 
> 
> Kind Regards,
> 
> Hong Sun | Agfa HealthCare
> Researcher | HE/Advanced Clinical Applications Research
> T  +32 3444 8108
> 
> http://www.agfahealthcare.com
> http://blog.agfahealthcare.com
> Click on link to read important disclaimer: http://www.agfahealthcare.com/maildisclaimer 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20130307/0d9714ea/attachment-0001.html>


More information about the bioontology-support mailing list