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] Request [bioontology-info] Mappings

Clement Jonquet jonquet at stanford.edu
Wed Nov 25 10:16:38 PST 2009


Dear Robin, 

Yes indeed, your request is much more clear to me know. Thank you for having clarified it.
So:

OBA_v1.1_documentation.htm is indeed an old documentation that refers the first Annotator prototype... I need to update this the wiki.

Indeed, we are using locally the workflow of the Annotator to index per ontology concepts biomedical resources. PubChem is one of these resources.
You will have more information about how to query the "NCBO Resource Index" here:
http://www.bioontology.org/wiki/index.php/Resource_Index 

Check: http://rest.bioontology.org/resource_index/resources/PCM 
inorder to see the fields that we have indexed for PCM resource.

Note that when the annotatiosn are created we now use a by default stop word list available here:
http://rest.bioontology.org/obs/stopwords
(however, as it has been implemeted after some executions, there might be some annotations created with stop words yet in the index).

For now, at query time, there is no possibility to filter the annotations using you own list of stop words. This is is currently being implemented.

Concerning your issue to figure out which ontology to use. Your request is indeed very justified.
For now, we do not have a service API that allows to get the statistics (i.e., number of annotations) per ontology. But as this request was made several time recently, I have submitted a feature request to implement it. I will let you know.

Also, I would like to mention you another service that we are developping and that could help you figure out which ontology to use for your data: 
http://stage.bioontology.org/recommender
This service will use the Annotator with all the availble ontologies and rank the ontologies by order of coverage. 

Hope it helps, 
Let me know.

Clement


----- Original Message -----
From: "Robin Smith" <RSmith at med.miami.edu>
To: "Clement Jonquet" <jonquet at stanford.edu>
Sent: Tuesday, November 24, 2009 7:21:54 AM GMT -08:00 US/Canada Pacific
Subject: RE: Request [bioontology-info] Mappings

Clement,

As it turns out, we were using the wrong server.  Somehow we ended up on this page:
http://obs.bioontology.org/docs/oba/OBA_v1.1_documentation.htm

Actually the cheminformatics database I was referring to is Pubchem, which it appears you have already indexed.  Is there any way in which we could explore that index outside of a simple web form?

What we'd like to do is find out which NCBO ontologies describe Pubchem the best, so that we can extract domain specific vocabulary to construct a new ontology.  Using the NCBO annotator, we're finding that some ontologies, such as the NCI thesaurus or SNOMED, have thousands of terms that annotate Pubchem.  However, a lot of these terms are not specific to any biotechnology domain (example Snomed terms: "For", "two", "one", Is a"....).  You could annotate any block of text from any source with these terms. If we knew how many "terms" each ontology was comprised of, then we might be able to correct for large ones to find which ones have a disproportionate number of terms that describe our data set. 

Hopefully I am making myself clear, and not misunderstanding anything - please let me know if you have questions.

Robin

-----Original Message-----
From: Clement Jonquet [mailto:jonquet at stanford.edu] 
Sent: Monday, November 23, 2009 3:01 PM
To: Smith, Robin
Subject: Request [bioontology-info] Mappings


Dear Robin, 

I am glad to hear that you have been using the Annotator for your data.
Please keep us informed of the interesting results or insights that our services would have help you in.

Let me be sure that you use the right version of the Annotator (i.e., the production server) because ncbolabs-dev2 is a developement machine that you shoud not be using.
Be sure to use the service described here:
http://www.bioontology.org/wiki/index.php/Annotator_Web_service


Concerning your request, could you detail a bit your thoughts:
> in which the total number of annotations for each ontology is listed.
- Annotations of what ?

We are also working on the NCBO Resource Index in which we used the Annotator to index biomedical reosurces:
http://www.bioontology.org/wiki/index.php/Resource_Index
Are you interested about the statistics in this index ?


Best regards, 

Clement

------------------------------------------------------------------------------------------------------------------------------

Dr. Clement JONQUET - PhD in Informatics

Postdoctoral Research Fellow - Stanford University - National Center for Biomedical Ontology

 

*email*:       jonquet at stanford.edu                                      *web*:      http://www.stanford.edu/~jonquet

*tel/fax*:    (001) 650 724-0388 / (001) 650 725-7944        *skype*:   clementpro

 

Stanford Center for Biomedical Informatics Research (BMIR)           http://bmir.stanford.edu/

Medical School Office Building, Room X-215

251 Campus Drive

Stanford, CA 94305-5479  USA

------------------------------------------------------------------------------------------------------------------------------


Begin forwarded message:


From: "Smith, Robin" <RSmith at med.miami.edu>
Date: November 23, 2009 8:58:46 AM PST
To: "'info at bioontology.org'" <info at bioontology.org>
Subject: [bioontology-info] Mappings


To Whom it May Concern,
 
My name is Robin Smith and I'm a postdoc at the University of Miami.  We've been using the NCBO Annotator web service recently to annotate a large cheminformatics database and have gotten some great data out of it. 
 
I was wondering whether you could provide us with an XML/tab delimited file,  similar to http://ncbolabs-dev2.stanford.edu:8080/OBS_v1/obs/ontologies , in which the total number of annotations for each ontology is listed.  We've been noticing that some ontologies, such NCI and SNOMED, provide almost too many annotations, and it would be helpful to be able to normalize to see which ontologies describe our data set the best.
 
Thanks,
 
Robin
 
_______________________________________________
bioontology-info mailing list
bioontology-info at lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/bioontology-info




More information about the bioontology-support mailing list