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] Query to get children.

Eamonn Maguire maguire at ebi.ac.uk
Fri Aug 28 02:25:26 PDT 2009


Hi Misha,

I will send my document by the end of the day detailing what we want  
to do and also showing the User interfaces we already have in place.

On the OBS issues, I realised this when Trish initially explained to  
me what happens in the back end. I knew that some sort of reasoning  
had to occur to get subclasses or parent classes (since this is how i  
build the trees to display the ontologies to the user inside the  
isacreator configuration tool), but i didn't realise just how out of  
sync these resources were, particularly with respect to OBI which OBS  
serves the oldest version of. Is building a simple service which  
reasons over the ontologies available each week, creating the data  
required for OBS out of the question?

I realise that I'm placing extra work on your shoulders, but i truly  
believe that such queries could be extremely useful. As mentioned in a  
previous email, i'm happy to contribute to the work since it probably  
aids us more than anyone else at this point in time. Perhaps we can  
discuss this in more detail when we come to meet you in two weeks time?!

Thanks again,

Eamonn

On 27 Aug 2009, at 20:01, Michael Dorf wrote:

> Hi Trish,
>
> I mean, any UI client.  I simply need to see a) how the search service
> itself needs to change, and b) the need for any potential new  
> services to
> satisfy the use-cases for this feature (for example, a dropdown that  
> lists
> all concepts for a given ontology etc...)
>
> Another relevant observation that somewhat complicates the  
> implementation
> of this feature is that the service that returns children for a  
> given term
> is currently implemented as an OBS wrapper, meaning it will use the  
> OBS
> database to compute the children hierarchy.  As you probably know,  
> there
> is a data discrepancy between BioPortal and OBS caused by the varying
> update frequency in two applications.  So there can easily be a case  
> of
> user searches of ontologies that either have not yet made it to OBS or
> have the latest indexed version in BP inconsistent with that in OBS.
>
> Thanks,
>
> Misha
>
>
> -----Original Message-----
> From: Trish Whetzel [mailto:whetzel at stanford.edu]
> Sent: Thursday, August 27, 2009 11:41 AM
> To: 'Michael Dorf'; maguire at ebi.ac.uk
> Cc: support at bioontology.org; 'Natasha Noy'
> Subject: RE: Query to get children.
>
> Misha, do you mean UI from BioPortal or UI from the ISAcreator tool?
>
> Eamonn, can you provide these additional details?
>
> Thanks,
> Trish
>
>> -----Original Message-----
>> From: Michael Dorf [mailto:mdorf at stanford.edu]
>> Sent: Thursday, August 27, 2009 10:26 AM
>> To: 'Trish Whetzel'; maguire at ebi.ac.uk
>> Cc: support at bioontology.org; 'Natasha Noy'
>> Subject: RE: Query to get children.
>>
>> Hi Trish,
>>
>> It would help to get a better picture of what type of UI workflow  
>> this
>> additional feature would require.  What would one need to do (type  
>> in,
>> select from a dropdown etc.) to have the search narrowed to a  
>> specific
>> branch.  I presume the user would need to know the exact name of the
>> concept that will serve as the top of the tree. Would the user just  
>> type
>> it into a textbox? Will the search include sibling concepts and their
>> sub-trees? What happens if the user-entered concept doesn't exist?
>> Would
>> we expect a full-concept including the namespace or a short one?
>>
>> In short, I do need more detailed requirements to produce an  
>> estimate.
>> A
>> couple of use-cases with success and exception scenarios would be  
>> ideal.
>>
>> Thanks!
>>
>> Misha
>>
>>
>>
>> -----Original Message-----
>> From: Trish Whetzel [mailto:whetzel at stanford.edu]
>> Sent: Wednesday, August 26, 2009 12:18 PM
>> To: maguire at ebi.ac.uk
>> Cc: 'Michael Dorf'; support at bioontology.org
>> Subject: RE: Query to get children.
>>
>> Got it.
>>
>> Misha, are there more details that are needed to get a time  
>> estimate to
>> implement this?
>>
>> Trish
>>
>>> -----Original Message-----
>>> From: maguire at ebi.ac.uk [mailto:maguire at ebi.ac.uk]
>>> Sent: Wednesday, August 26, 2009 12:05 PM
>>> To: Trish Whetzel
>>> Cc: 'Eamonn Maguire'; 'Michael Dorf'; support at bioontology.org
>>> Subject: RE: Query to get children.
>>>
>>> hi trish,
>>>
>>> the second case is the correct one :o)
>>>
>>> thanks,
>>> Eamonn
>>>
>>>> To clarify the request a bit further, would the parameter
>>> 'searchBranch'
>>>> be
>>>> predefined terms in various ontologies? For example, the set of  
>>>> terms
>>> that
>>>> can be used to annotate various object models such as FuGE, MAGE- 
>>>> ML,
>>> etc.
>>>>
>>>> OR
>>>>
>>>> Would 'searchBranch' be a user-defined term and would be used to
>> limit
>>> the
>>>> search space to only those terms that are children terms of the  
>>>> term
>>>> specified in 'searchBranch'?
>>>>
>>>>
>>>>
>>>> Trish
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> From: Eamonn Maguire [mailto:maguire at ebi.ac.uk]
>>>> Sent: Friday, August 21, 2009 10:11 AM
>>>> To: Michael Dorf
>>>> Cc: 'Trish Whetzel'; support at bioontology.org
>>>> Subject: Re: Query to get children.
>>>>
>>>>
>>>>
>>>> Hi Michael,
>>>>
>>>>
>>>>
>>>> This is very useful.
>>>>
>>>>
>>>>
>>>> Coming back to the original feature request of adding an extra
>>> parameter
>>>> to
>>>> the search query to specify a branch to search under, how long do  
>>>> you
>>>> think
>>>> it will take before this feature is implemented? By the way, I'm
>> happy
>>> to
>>>> contribute with the effort of doing it since any implementation I  
>>>> do
>>> now
>>>> will simply be replaced by whatever is created by you guys when the
>>> actual
>>>> ability to query under a branch is realised.
>>>>
>>>>
>>>>
>>>> Any code i write now will essentially take the result of the
>> hierarchy
>>>> function and then filter for terms containing the string searched  
>>>> for
>>> by
>>>> the
>>>> user.
>>>>
>>>>
>>>>
>>>> Thanks,
>>>>
>>>>
>>>>
>>>> Eamonn
>>>>
>>>>
>>>>
>>>> On 20 Aug 2009, at 19:42, Michael Dorf wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hi Eamonn,
>>>>
>>>>
>>>>
>>>> This is the correct URL to use:
>>>>
>>>>
>>>>
>>>> http://rest.bioontology.org/obs_hibernate/ontologies
>>>>
>>>>
>>>>
>>>> The complete list of OBS services is available here:
>>>>
>>>>
>>>>
>>>> http://rest.bioontology.org/test_obs.html
>>>>
>>>>
>>>>
>>>> The Hierarchy Services subsection of this page is the one you are
>>>> referencing from BioPortal. For example, the service
>>>>
>>>>
>>>>
>>>> http://rest.bioontology.org/obs_hibernate/parents/MSH/C0025202
>>>>
>>>>
>>>>
>>>> is also available as
>>>>
>>>>
>>>>
>>>> http://rest.bioontology.org/bioportal/concepts/parents/MSH/C0025202
>>>>
>>> <http://rest.bioontology.org/bioportal/concepts/parents/MSH/C0025202%20
>>>
>>>> in
>>>> a slightly different (more BioPortal-centric) XML format.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hope this helps.
>>>>
>>>>
>>>>
>>>> Misha
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> From: Eamonn Maguire [mailto:maguire at ebi.ac.uk]
>>>> Sent: Thursday, August 20, 2009 9:39 AM
>>>> To: Trish Whetzel
>>>> Cc: Michael Dorf; support at bioontology.org
>>>> Subject: Re: Query to get children.
>>>>
>>>>
>>>>
>>>> I can't access it no.
>>>>
>>>>
>>>>
>>>> Eamonn
>>>>
>>>> On 20 Aug 2009, at 17:35, Trish Whetzel wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> The hierarchy services are currently geared towards certain  
>>>> versions
>>> of an
>>>> ontology, so 39885 is the version that has been processed to  
>>>> generate
>>> data
>>>> for the ontology hierarchy services. This version also includes the
>>> 'efo:'
>>>> prefix for the term while the version 40394 does not.
>>>>
>>>>
>>>>
>>>> There is a service to call to get the list of ontologies that are
>> used
>>> in
>>>> the ontology hierarchy service, however I have been having  
>>>> difficulty
>>>> accessing this from outside Stanford and trying to sort out the
>>> reason.
>>>> The
>>>> URL is: http://ncbo <http://ncbo/>
>>>> -obs-prod1:8080/obs_hibernate/ontologies.
>>>> Are you able to access this?
>>>>
>>>>
>>>>
>>>> Trish
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Aug 20, 2009, at 9:13 AM, Eamonn Maguire wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hi trish et al :o)
>>>>
>>>>
>>>>
>>>> now, this is closer to what i need :o) i didn't realise this  
>>>> service
>>> was
>>>> available! if anything there is too much metadata associated with
>> this
>>>> query
>>>> though. We'd only really metadata similar to that returned in
>>> bioportals
>>>> search function. For a large hierarchy, this file could be massive
>>>> particularly in an ontology with a hierarchy containing large  
>>>> numbers
>>> of
>>>> axioms. You'd essentially be returning a subset of the OWL file  
>>>> in a
>>>> differing XML format.
>>>>
>>>>
>>>>
>>>> however i see that the version of EFO you are making me query  
>>>> with is
>>> old.
>>>> Whenever i attempt to query with the latest version i get the same
>>> error
>>>> that i got yesterday when i reported the problems with the children
>>> query
>>>> returning a 'No child in hierarchy found' message.
>>>>
>>>>
>>>>
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/children/40394/efo:EFO_0
>> 0
>>> 0035
>>>> 2
>>>>
>>>>
>>>>
>>>> or
>>>>
>>>>
>>>>
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/40394/efo:EFO_0001456?em
>> a
>>> il=e
>>>> xample at example.org
>>>>
>>>>
>>>>
>>>> Thanks,
>>>>
>>>>
>>>>
>>>> Eamonn
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On 20 Aug 2009, at 16:49, Trish Whetzel wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hi Eamonn,
>>>>
>>>>
>>>>
>>>> The last I talked to the developers on the design of the bean from
>> the
>>>> hierarchy services the information I received is that the hierarchy
>>>> service
>>>> needs to be used with the concept service, e.g.
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/39885/efo:EFO_0000352?em
>> a
>>> il=e
>>>> xample at example.org, to generate the additional term information.
>>>>
>>>>
>>>>
>>>> Misha, can you add more on this design decision and whether the
>>> preferred
>>>> term field can be added to the bean from the ontology hierarchy
>>> service?
>>>>
>>>>
>>>>
>>>> Thanks,
>>>>
>>>> Trish
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Aug 20, 2009, at 3:58 AM, Eamonn Maguire wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Thanks trish,
>>>>
>>>>
>>>>
>>>> So now my next question relates to the metadata. In the result from
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/children/39885/efo:EFO_0
>> 0
>>> 0035
>>>> 2 we have a result which looks like this:
>>>>
>>>>
>>>>
>>>> <classBean>
>>>>
>>>>    <ontologyVersionId>39885</ontologyVersionId>
>>>>
>>>>    <id>efo:EFO_0000493</id>
>>>>
>>>>    <relations>
>>>>
>>>>        <entry>
>>>>
>>>>            <string>Level</string>
>>>>
>>>>            <int>1</int>
>>>>
>>>>        </entry>
>>>>
>>>>    </relations>
>>>>
>>>> </classBean>
>>>>
>>>>
>>>>
>>>> now, this doesn't give me much in terms of anything i can use
>> straight
>>>> away.
>>>> can i request the addition of the concept name and concept short  
>>>> name
>>> like
>>>> you have in the main search result to avoid making many additional
>>> queries
>>>> to find out what EFO_0000493 actually is.
>>>>
>>>>
>>>>
>>>> For those of you in this email who don't know the use-case we  
>>>> want to
>>>> implement, i'll give you a quick overview.
>>>>
>>>>
>>>>
>>>> inside isacreator we offer people the choice of querying either a
>>>> recommended ontology (provided through a configuration tool) or to
>>> search
>>>> the whole ontology. the latter case is already fulfilled by  
>>>> BioPortal
>>>> obviously :o) in the case of querying a recommended ontology(ies)  
>>>> you
>>> also
>>>> have what we would need if we just wanted to query for terms  
>>>> matching
>>>> 'protocol' for example in OBI or EFO through supplying ontology  
>>>> ids.
>>>> However, what if i wanted to restrict the search space as well as  
>>>> the
>>>> ontology, meaning restricting the search to a certain part of one  
>>>> or
>>> more
>>>> ontologies. e.g., in EFO
>>> http://bioportal.bioontology.org/visualize/40394
>>>> when a search is made, only the terms under 'process (EFO_0001433)'
>>> should
>>>> be searched for.
>>>>
>>>>
>>>>
>>>> I have already requested via trish an extra parameter in BioPortals
>>> search
>>>> query to allow one to provide an ontology term id or unique name
>>>> (dependent
>>>> on the ontology being used) to restrict the search space to a
>>> particular
>>>> part of the ontology.
>>>>
>>>>
>>>>
>>>> the query would look like this:
>>>>
>>>>
>>>>
>>>>
>>> http://rest.bioontology.org/bioportal/search/protocol/?ontologyids=3988
>> 5
>>>>
>>> <http://rest.bioontology.org/bioportal/search/protocol/?ontologyids=398
>> 8
>>> 5&se
>>>> archBranch=> &searchBranch=EFO_0001433
>>>>
>>>>
>>>>
>>>> not only would such a search be greatly advantageous to us, but it
>>> could
>>>> also speed up searches made by others on bioportal since the search
>>> space
>>>> is
>>>> restricted!
>>>>
>>>>
>>>>
>>>> i hope this is clear,
>>>>
>>>>
>>>>
>>>> thanks for all your great efforts thus far!!
>>>>
>>>>
>>>>
>>>> Eamonn
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On 19 Aug 2009, at 17:26, Trish Whetzel wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> There was a change in the identifiers for EFO a few versions back.
>> The
>>>> change is that there was an extra 'efo' in the fully qualified term
>>> name.
>>>>
>>>>
>>>>
>>>> The term id in version 39885 is:
>>>> http://bioportal.bioontology.org/visualize/39885/efo:EFO_0000001
>>>>
>>>> The term id in version 40394 is:
>>>> http://bioportal.bioontology.org/visualize/40394/EFO_0000001
>>>>
>>>>
>>>>
>>>> So using
>>>>
>>> http://rest.bioontology.org/bioportal/virtual/children/1136/efo:EFO_000
>> 0
>>> 352,
>>>> will get the child term of clinical history or
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/children/39885/efo:EFO_0
>> 0
>>> 0035
>>>> 2 will work.
>>>>
>>>>
>>>>
>>>> Trish
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Aug 19, 2009, at 6:16 AM, Eamonn Maguire wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hi Trish,
>>>>
>>>>
>>>>
>>>> Just a quick question.
>>>>
>>>>
>>>>
>>>> I'm having trouble getting the "Get parents/children of a given
>>> concept in
>>>> a
>>>> specific ontology version" query to work properly. For instance, I
>>> thought
>>>> that making the following queries should work, but they return a
>>> message
>>>> saying that no children exist in the hierarchy when they clearly  
>>>> do.
>>>>
>>>>
>>> http://rest.bioontology.org/bioportal/concepts/children/40394/EFO_00003
>> 5
>>> 2
>>>>
>>>>
>>>>
>>>> i get the same message using the virual ontology equivalent of this
>>> method
>>>>
>>>>
>>> http://rest.bioontology.org/bioportal/virtual/children/1136/EFO_0000352
>>>>
>>>>
>>>>
>>>> any ideas?
>>>>
>>>>
>>>>
>>>> thanks,
>>>>
>>>> Eamonn
>>>>
>>>> --
>>>>
>>>> *    maguire at ebi.ac.uk
>>>>
>>>> :     http://www.ebi.ac.uk/~maguire
>>>>
>>>> :     http://www.ebi.ac.uk/net-project
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> *    maguire at ebi.ac.uk
>>>>
>>>> :     http://www.ebi.ac.uk/~maguire
>>>>
>>>> :     http://www.ebi.ac.uk/net-project
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> *        maguire at ebi.ac.uk
>>>>
>>>> :         http://www.ebi.ac.uk/~maguire
>>>>
>>>> :     http://www.ebi.ac.uk/net-project
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> *            maguire at ebi.ac.uk
>>>>
>>>> :             http://www.ebi.ac.uk/~maguire
>>>>
>>>> :     http://www.ebi.ac.uk/net-project
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> *          maguire at ebi.ac.uk
>>>>
>>>> :           http://www.ebi.ac.uk/~maguire
>>>>
>>>> :     http://www.ebi.ac.uk/net-project
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>

--
* 	maguire at ebi.ac.uk
: 	http://www.ebi.ac.uk/~maguire
:     http://www.ebi.ac.uk/net-project










-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20090828/b1f921cb/attachment.html>


More information about the bioontology-support mailing list