[Data-modeling] Thoughts on disease/treatment

[Benjamin Good]

They do indeed have and OWL version of this anatomical beast, though  
its native representation is Protege frames.  There is a link to a  
"lite" version of this ontology here  http://obofoundry.org/cgi-bin/detail.cgi?id=fma_lite

I mention this here on the main list rather than in the Medicine  
discussion forum because its yet another example of a problem that I  
think everyone in the semantic web community that is interested in  
freebase (which really should be everyone) really needs an answer to.   
What is a good protocol for moving OWL ontologies into and out of  
freebase ?  Can this problem be dealt with in a domain/ontology  
independent manner?

I am facing this problem directly in my work on the entity describer (www.entitydescriber.org 
).  One of the greatest selling points I've had within the  
bioinformatics community is that I can use it as a generic Gene  
Ontology (GO) annotation service by pointing the topic lookup at the  
GO group type (http://www.freebase.com/view/biology/ 
gene_ontology_group) (and that, as an indication of the already  
extensive community, someone unknown to me had already imported the  
GO).  However, there are a number of problems:

1) the GO import could be made much useful through better integration  
with the rest of freebase.  For example, http://www.freebase.com/view/en/biological_reproduction 
  should very likely be linked to GO:reproduction (http://www.freebase.com/view/guid/9202a8c04000641f800000000520ead4 
).  How might this be automated ?

2) is there a mechanism to try to update the freebase data when the GO  
is updated ?
3) users of things like the GO really expect and depend on inferences  
- especially hierarchical ones (isa, partof).  What is a good pattern  
for representing hierarchy and executing queries across hierarchies  
within the context of freebase?

For the entity describer project we are currently debating whether to  
establish (or find) a queriable repository of ontologies like the GO  
for the purpose of answering queries that require inference or trying  
to get everything done in the context of freebase.

-Ben








On Jun 6, 2008, at 2:03 PM, Dan Ruderman wrote:

> Hi Faye,
>
> Faye Li wrote:
>> I have a to-do item regarding anatomical structures affected by  
>> diseases
>> on my whiteboard but didn't get to it this time. The type "Anatomical
>> structure" exists today without any properties (see
>> http://www.freebase.com/tools/schema/medicine/anatomical_structure)  
>> and
>> the plan was to try to flesh out the schema there. Any properties  
>> you'd
>> suggest?
>>
> The fundamental concept I would add is a
> contains/part-of hierarchy to the structures.  I found
> this web site which has thought a lot of this through and
> may have an ontology which is importable (e.g. they
> have an OWL version):
> http://sig.biostr.washington.edu/projects/fm/AboutFM.html
> Populating the hierarchy might be tricky, though, unless
> the anatomical names are consistent with the freebase ones.
> My biggest concern would be that the hierarchy would be
> different in different organisms (e.g. cerebral cortex is part of the
> mammalian brain but not part of the reptilian brain), and
> I'm not sure if there is a ready source for that information.
> Because of these complexities I'd probably leave the properties
> as they are.
>
>
>> As for "Disease Cause" (which will be renamed etiology shortly,  
>> with the
>> original name saved as an alias), I was thinking about refactoring  
>> the
>> type. Each cause needs to be qualified with evidence level, something
>> along the lines of "evident, probable, possible". I was also  
>> considering
>> adding an enumeration property/new type for etiology category that  
>> would
>> list, "bacterial, viral, chemical, parasitic", etc. I would  
>> appreciate
>> your expertise if you have time to talk about this offline.
>>
> My knowledge is somewhat limited, but I enjoy thinking about
> these things and would be glad to chat about them.
>
> Dan
>
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling