I again apologize for having missed this important discussion due to conflicting commitments.
I wanted to add just a few brief comments to the discussion points listed in the minutes:
What is the goal of BIONT? (DD)
Goal is not a demo per se. It's to create artifacts listed on the BIONT wiki. (VK)
BB: Specifically targeting support for the Use Cases being pursued collectively by all the Task Forces, correct? If so, a demo that includes BioONT efforts is implied, since the Use Cases will lead to a demo.
When we say best practices developing ontologies...isn't NCBO doing this? (TC)
NCBO's goal is to provide tools and technologies to support the collaborative process. (VK)
Exactly. And they have as their mandate to be maximally inclusive both in terms of artifacts they distribute to the community and practices pursued across the community. Mark Musen discusses some of the complexities of dealing with these differing views & practices in biomedical ontology development in his presentation from this 2005 Biomedical Ontology-driven Informatics meeting from 2005 (http://ontology.buffalo.edu/wg6/
), as do the presentation by several others at this meeting (e.g., Chris Chute [http://mayoresearch.mayo.edu/mayo/research/staff/chute_cg.cfm]
from Mayo Clinic whose LexGrid system [http://informatics.mayo.edu/LexGrid/index.php?page=tnt] needs to support this fact in the context of providing lexicon+ontology support for the NCI caBIG project, as does their related Open Terminology Services (OTS) project [http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=14728514&query_hl=1&itool=pubmed_docsum] which I believe has been absorbed into LexGrid as their implementation of the HL7 Common Terminology Service [http://informatics.mayo.edu/LexGrid/index.php?page=projects]).
Having said this, the OBO Foundry IS promoting a collection of best practices which appear to have been distilled from the experience of the Gene Ontology Consortium, the Stanford Medical Informatics group, the IFOMIS center's efforts in regards to the various applications to which the BFO foundational ontology has been put in the biomedical domain, and several other community biomedical concept-mapping, classification, and ontology development efforts. I do think it is extremely important we stay abreast of those ontologies being collected in the OBO Foundry (e.g., BFO, PATO, OBI, etc.) and the various applications where they are being used.
Look at experiences of the GO project. (VK)
BB: see above comment - plus I would strongly recommend taking a look at the slides contributed by Suzi Lewis & Michael Ashburner (two of the GO founding PIs) to the presentation "Principles for Building Biomedical Ontologies" listed at the GO educational resource page (http://www.geneontology.org/GO.teaching.resources.shtml?all
Is this more of a group self help effort? (TC)
BB: I'm not certain I understand what Tim means here but would love to know more about the discussion that took place regarding this comment.
We could compile experiences integrating data using existing open source ontologies such as FMA. (VK)
I have an Anatomical Informatics bibliography I've compiled over the years that includes nearly all the applications to which FMA has been applied of which I'm aware - along with other related anatomical ontology efforts such as those associated with the Edinburgh Mouse Atlas Project & the JAX/MGI/GO effort, if that could be of some use on this point.
start to put together data/ontology we have now, to see waht query/isses we can get out of these data
Exactly - to my mind, this is THE critical task BioONT needs to invest in - essentially the iterative process of honing ontology support for the Use Cases.
ACTION: Map ontologies to data sources (VK)
BB: see some of the issues I bring up in the Word file I've attached below
ACTION: Have teams with experience using ontologies to integrate data talk about...
BB: Great idea!
ACTION: get feedback from various group on the issues that came up in developing ontologies and mapping to data
BB: I'd be glad to provide feedback based on ontology development & mapping data based on our BIRNLex work within the BIRN project.
Attached please find some very detailed descriptions of how to fully flesh out the required HCLSIG ontologies derived from our recent communal manuscript editing effort. None of this made it into the final draft (where it really didn't belong anyway), so many may have not seen it all, but I would suggest we seek this level of due diligence to pull together references to the relevant, extant public knowledge sources covering the required domains.
The idea here is in order to be maximally useful to the BioRDF effort, the ACPP effort, the efforts in the publishing task force, and other tasks forces as they pursue supporting the Use Cases, we should try to link to all the unique, community resources relevant to the entities and relations we need to cover. I would also want to spend more time reviewing this with Matthias to determine how many of these relevant links can be derived directly from his bio-zen work, as well as discussing with Kei and others within BioRDF the ways in which they may have already established some of these links.
As an additional thought to consider on this issue of supporting ontologies, I would suggest we consider pursuing a manuscript to describe how we've been approaching this task - one highlighting the BioONT ontology artifacts and our means of incorporating appropriate, extent community ontologies and related knowledge resources. It would center around the Use Cases and describe some of the issues that have come up as we've worked to support the formal semantic framework/ontology requirements for all of the other HCLSIG Task Forces. Pursuing such a manuscript would probably assist in further specifiying the BioONT commitment to all the other task forces, as well.