I received funding for a Collaborative Exchange Opportunity through the Phenotype RCN and I visited Chris Mungall at Berkeley two weeks ago. I am a member of the group working on the FEED Database project (http://www.feedexp.org/wiki/Public:Feeding_Experiments_End-User_Database), which began as a NESCent working group and is now funded directly by NSF. The aim of the FEED project is to create an online database that will act as a repository for physiologic and kinematic data on feeding and other behaviors of mammals. As part of the project we are working to create definitions for musculo-skeletal structures of the head and neck and for behaviors of the oro-pharyngeal complex that will work for all eutherians and metatherians. At a meeting in May, we created the spreadsheet of definitions that some of you may have seen on Phenoscape curator listserve and another one for oro-pharyngeal behaviors. As a newcomer to the field of ontology, I will be attending the course at NESCent at the end of the month, but I am anxious to continue the work we began in May so I met with Chris to figure out how to proceed. I was very happy with what we accomplished in a very short time. Here is a brief summary of our work:
* We looked at Phenote as an alternative to using an Excel spreadsheet for curating muscle properties, and decided this might work quite well. The Phenote file should be checked into phenoscape svn somewhere, but it is probably a bit premature to do this. The resulting properties can then be automatically merged into an ontology.
* We added the current spreadsheet FEED definitions into Uberon using the external_definition annotation property. See for example – http://purl.obolibrary.org/obo/UBERON_0001597 (it doesn’t look perfect in OntoBee yet, but soon it will have a link to feedexp.org).
* When the FEED group met and created the spreadsheet of definitions we realized very quickly that any definition of a muscle that is applicable to all mammals must be very generic. This takes a lot of “getting used to” for anatomists (for an example, take a look at our recent exchange regarding the genioglossus muscle on the Phenoscape curators listserve). We have found that the definitions of the head and neck muscles that we have looked at in Uberon were taken from dbpedia and are specific to human anatomy. These definitions must be revised to be useful for comparative anatomical studies.
* As a first try, we looked at some of the attachment sites already in Uberon taken from dbpedia and improved some of them (e.g. buccinator).
* I would like to be able to create character state matricies from our annotated database so we attempted to use Phenex to annotate the character state matrix in my paper from a recent FEED symposium (Druzinsky et al 2011) paper but we didn’t get very far. We thought it would be useful to cover Phenex in the upcoming course.
* We did a whirlwind Protege tour including some fairly complex features such as axiom annotations (i.e. per-statement attribution).
* We looked briefly at NBO and at the behavior spreadsheet. I am going to contact some of the FEED collaborators to investigate the possibility of getting these into NBO.
Robert Druzinsky (U. of Illinois, Chicago) with lots of help from Chris Mungall (U.C. Berkeley) and thanks to Heiko Dietze