biology Topics
-
-
Hi :-)
I'm a grad student, and I'm helping to develop an experiment which will ask people to annotate web services with tags. The experiment is part of the ED project, described at http://www.connotea.org/wiki/EntityDescriber.
What I want to do is load the list of candidate tags into Freebase as Topics, all under a single Type. However, there are a lot of them. (I estimate about 100,000.) I expect that I will hit the user upload limit when I try to do this, and I was wondering if you could raise the limit temporarily for me? (50,000 assertions per day would be okay, and it could be set back to normal in two weeks.)
Thanks for your help,
- Ben Vandervalk
Masters Student
University of British Columbia
Wilkinson Lab
-
Hi Ben,
I've passed your request on to the data team - someone should be replying to you shortly.
-
Hey Ben -
Are working with Benjamin Good on this? I thought his original idea for using Freebase to annotate Connotea was great. I am very supportive of using Freebase IDs as content tags.
Could you send me a note about the nature of the candidate tags. Are they tied to other source vocabularies we could reference? I'd also be curious if there are other semantic relationships we could draw out between the tags (or other Topic vocabularies) to make navigating the annotated relationships in Connotea more powerful.
Jamie [at] metaweb
-
Hi Jamie,
Wow, you know about ED already. Cool!
Yes, I am working with Ben Good on this experiment. Basically, what he is doing is using ED to annotate BioMOBY web services with Freebase Topics (each BioMOBY web service has a unique URI). In addition to general tags to describe a service, he is also collecting a special set of tags which describe the relationship between the input and output data: i.e. "Each output is a about/of the input(s)", e.g. "Each output is a _Homolog_ about/of the input(s)"
Since the services are all bioinformatics related, my plan was to load terms from a biology/bioinformatics ontology into Freebase. My choice is the NCI Thesaurus (http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl) , which has about 80,000 classes representing all sorts of biological things.
The main reason we want to use Freebase for this is because of the great type-ahead feature. I was planning to keep it simple and just load all the terms under one Type. So to be honest, it's "throw away" data, that I would remove from Freebase after the experiment was done.
Thanks for your interest :-)
- Ben Vandervalk
-
Oh also, you can browse the NCI Thesaurus at http://bioportal.bioontology.org/visualize/13578
- Ben V.
-
Hi Ben,
I've been working on some ontology mapping types based off of some of Jamie's own Web Ontology types. If you import 80,000 topics into Freebase I would be really interested in seeing those Freebase topics mapped back to their original URIs in the NCI Thesaurus.
To give you some idea of how this would look, I have a basic Freebase to FOAF mapping here.
Shawn
-
Cool, nice work :-) That's the proper way to do it, I think.
My intention is to something quick and dirty though, just so I can use the Freebase type ahead feature in our annotation experiment. Mapping all 80,000 classes and umpteen properties to existing Freebase entities would be a ton of work, and I just don't have time to do that!
- Ben V.
-
Hi Ben,
Have you been communicating with Jamie about this? Can you send him an email regarding the nature of the candidate tags - his email is jamie [at] metaweb. Thanks.
-
Sure, I'll copy my message above to jamie@metaweb.com.
-Ben V.
-
