Re: Comparing neurolex against an ontology

[Stephen Larson]

As the documentation page says:

"The RDF is refreshed every hour both as a file and within the endpoint in order to facilitate changes on the wiki being usable for external SPARQL reasoning in near-realtime."

External triple store: 

http://jena.apache.org/documentation/serving_data/

Thanks,

  Stephen

P.S. There's a google group for NeuroLex: https://groups.google.com/forum/#!forum/neurolex

On Tue, Jul 30, 2013 at 11:44 AM, Trish Whetzel <whe...@ucsd.edu> wrote:

Is information from the wiki added immediately to the external triple store or if not, what is the time lag for synchs? What is the external triple store, Virtuoso, Owlim, etc?

Trish 

On Jul 30, 2013, at 6:22 AM, Stephen Larson <sla...@ncmir.ucsd.edu> wrote:

The documentation page that Anita pointed to should have all the information.  It is being served out of an external triple store, not the native one.

Thanks,

  Stephen

On Tue, Jul 30, 2013 at 9:16 AM, Maryann Martone <mar...@ncmir.ucsd.edu> wrote:

Stephen needs to weigh in there as well.  I thought we had ported the Neurolex into a triple store, because the native sparql endpoint wasn't very good.

Maryann Martone, Ph. D.
Professor-in-Residence
Department of Neuroscience
University of California, San Diego
San Diego, CA 92093-0446

858-822-0745

On Jul 29, 2013, at 3:24 PM, anita bandrowski <aband...@ucsd.edu> wrote:

> This is a good place to start.
>
> http://neurolex.org/wiki/NeuroLex_SPARQL_endpoint
>
> The end point and some documentation are there.
>
> On Mon, Jul 29, 2013 at 1:50 PM, Fahim Imam <smti...@gmail.com> wrote:
>> I am forwarding your question to Amarnath and Stephen. I think they should
>> know about the SPARQL interface for NLX.
>>
>>
>> ---------- Forwarded message ----------
>> From: Chris Mungall <cjmu...@lbl.gov>
>> Date: Mon, Jul 29, 2013 at 4:46 PM
>> Subject: Re: Comparing neurolex against an ontology
>> To: Trish Whetzel <plwh...@gmail.com>
>> Cc: Fahim Imam <smti...@gmail.com>
>>
>>
>> I am highly ignorant of semantic mediawiki technology. Apparently there is
>> some menu somewhere for a SPARQL query interface, so we could do a SELECT *
>> WHERE {?s ?p ?o} to get us started, but I haven't found this yet. I assume
>> someone somewhere in NIF is doing something involving the RDF form of NL?
>>
>>
>> On Mon, Jul 29, 2013 at 1:03 PM, Trish Whetzel <plwh...@gmail.com> wrote:
>>>
>>> These ideas generally sound the same, with the difference being at what
>>> step an OWL file is generated. Bubastis (http://www.ebi.ac.uk/efo/bubastis/)
>>> is a good and user friendly ontology diff tool. I'd have to review more
>>> deeply or ping James to know if it will identify the types of changes that
>>> can occur to a term through the NeuroLex wiki.
>>>
>>> Chris - do you know the wiki technology well enough to create the OWL
>>> file?
>>> Fahim - do you have a list of the types of changes one can make to a term
>>> from the wiki. I know adding synonyms is one type of change.
>>>
>>> Trish
>>>
>>>
>>> On Mon, Jul 29, 2013 at 12:43 PM, Chris Mungall <cjmu...@lbl.gov> wrote:
>>>>
>>>> My preference would be to separate out the functionality:
>>>>
>>>> 1. export OWL from NeuroLex (may actually be two steps - generate native
>>>> semantic wiki RDF and then translate this to the correct OWLTBox
>>>> representation)
>>>> 2. OWLDiff between nl.owl and relevant part of NIFSTD
>>>>
>>>> Note there are standard tools for #2 - e.g. James Malone's Boubastis. We
>>>> should re-use rather than re-invent here.
>>>>
>>>> If it runs against the OQ database it may limit it's applicability for
>>>> ontologies such as Uberon and GO.
>>>>
>>>> On Mon, Jul 29, 2013 at 12:20 PM, Fahim Imam <smti...@gmail.com> wrote:
>>>>>
>>>>> Hi Chris,
>>>>> It's mostly manual so far, unless there is a spreadsheet version
>>>>> available for a properly revised, curated list of terms (categories) in
>>>>> NeuroLex.
>>>>>
>>>>> As for the comparison there exist no straight-forward tool as of this
>>>>> moment. However, I was talking to Trish the other day about a possibility of
>>>>> a version comparison script that could run between neurolex and ontoquest
>>>>> database. These could be tracked through the 'Version ID' for each of the
>>>>> NeuroLex categories for a given range of time. Let me get back to you after
>>>>> I talk to Trish and Amarnath about the plan.
>>>>>
>>>>> Thanks,
>>>>> Fahim
>>>>>
>>>>>
>>>>> On Fri, Jul 26, 2013 at 7:41 PM, Chris Mungall <cjmu...@lbl.gov>
>>>>> wrote:
>>>>>>
>>>>>> What tools do you use?
>>>>>>
>>>>>> How do you export nlex as OWL?
>>>>>
>>>>>
>>>>
>>>
>>
>>
>
>
>
> --
> Anita Bandrowski, Ph.D.
> NIF Project Lead
> UCSD 858-822-3629
> http://neuinfo.org
> 9500 Gillman Dr.#0446
> la Jolla, CA 92093-0446
>

[Stephen Larson]

Glad you are finding the RDF dump useful.

No translation of that kind has been done -- would be useful to give it a shot!

We have a manuscript in submission for NeuroLex that explains a lot more of the modeling decisions.  We're hoping it is out soon.  That said, the model also evolves over time.  Any specific questions you have would be ideal for this list.

Thanks,

  Stephen

On Wed, Jul 31, 2013 at 3:51 PM, Chris Mungall <cjmu...@lbl.gov> wrote:

Thanks. I downloaded the RDF dump so I didn't need the SPARQL endpoint.

Is there already an implementation of a standard translation of this to an OWL TBox representation using standard class URIs?

Specifically

 - map URIs
   * http://neurolex.org/wiki/UBERON-3A ==> http://purl.obolibrary.org/obo/UBERON_
   * http://neurolex.org/wiki/CHEBI ==> http://purl.obolibrary.org/obo/CHEBI_
   * http://neurolex.org/wiki/Birnlex_ ==> corresponding NIFSTD URI (not clear if there is a clean way of doing this?)

 - map Datatype assertions on Property-3ALabel ==> Annotation assertions on rdfs:label

 - map Object Properties assertions such as partOf, OverlapsWith to existential class axioms using standard BFO or RO relation

 - change certain rdf:types to owl SubClassOf

I don't want to re-implement this if it's been done already. If I do, is there any document describing the modeling in neurolex? I find myself a little confused trying to determine this myself - for example, sometimes the rdfs:label contains an ID, sometimes it contains the actual label.