[seek-taxon] SEEK-Taxon Call Wednesday 6-14-06

[Beach, James H]

SEEK Taxon:
 
(Matt -- Any comments you might have here on the Kepler use case scenarios (see the bottom) would be useful.)
 
SEEK Taxon call tomorrow at 1600 GMT.
 
http://www.timeanddate.com/worldclock/fixedtime.html?month=6&day=14&year=2006&hour=16&min=0&sec=0&p1=0
 
 
The call in number will be the same as before:
 
(877) 278-8686  Conference ID/PIN: 653089
 
 
 
One item for the call tomorrow will be a discussion of the scope of work of a research assistant we will seek, to work with bat taxonomic concepts.  Before we let the advertisement out the door, I would like to have a clear idea of the tasks involved and how they relate to our ultimate objectives for the ENM use case with mammals of North America.  This will determine whether the person could work remotely or not, the amount of effort required, and duration of the appointment.
 
 
Here is a skeleton of some points for that scope of work for that person and for our discussion tomorrow:
 
 
Draft Tasks:
 
1) Use ConceptMapper (CM) to author additional taxon concepts for major bat treatments prior to MSW 1993 and those that may occur between 1993 and MSW 2005, and any significant new classifications after MSW 2005 was compiled.  This task would be aimed at obtaining as many bat classifications as practicable for the TOS (and all of their concepts).   Questions:  How many classifications?   All Sizes?  All bat groups?  Other Mammals also?   All levels of the hierarchy?  Big Question:  Does CM support the creation of concepts which INCLUDE children links in the definition?
 
2) Use CM to map among classifications using the notation for concept relationships, based at first only on published differences in the children belonging to the taxon.  This would be done for concepts at the genus level and above.  Are we referring to these as third-party annotations or are they something else?
 
3) Use CM to map among species concepts based on published, documented differences in characters, types?  Other kinds of data?  How easy will this be?
 
4) Grab several bat specimen databases from DiGIR providers, KU, and perhaps also from other institutions not in DiGIR, and query the TOS for the best concept for the species name on those specimens.  We might for example, just start doing this based on date of the specimen determination.  For simplicity, if a specimen was determined in 2006, and we have no other reference citation information on the label, then we might assign that name to the MSW concept of the same name for 2005.  If determined in 1994, we assign MSW93 concept to that name.
 
Assigning a concept to a determination would mean, querying the TOS for the best concept for that date of determination and name (again as a first simplifying step) and then getting an LSID from the TOS for that concept, and sticking it into the specimen record in the specimen database.  The specimen database that we use might start off as a slightly modified Darwin Core flat file that could accomodate LSIDs, or we might make something a little more normalized.  We can produce the software here that will take a specimen record and query the TOS for the best concept, and then automatically stick the LSID from the TOS concept into the specimen database.  Ideally we could do that in batch mode, for tens or hundreds of thousands of specimens, within a few minutes or hours.
 
That would give us multiple museum data sets for bats with LSIDs on their concepts.
 
------- Now the analysis part - for discussion (Matt take note) ----------
 
For Kepler, the DiGIR interface could now query bat species with the 'get best concept' service for any particular name, and return a concept for a particular classification for the name.  Right? Then we use the mappings in TOS to generate a list of mapped concepts, then use those (LSIDs) for querying our bat specimen datasets marked up with LSIDs. 
 
For the ENM use case, I believe the work flow would then be, to use all of the concepts from one classification to create a set of ENModels for those species, then choose a different classification and do the same, and then have some exciting way to contrast the results of the two sets of models, which hopefully will be using concepts of different definitions and sizes for at least some taxa.
 
 
Mañana.
_____________________________
James H. Beach
Biodiversity Research Center
University of Kansas
1345 Jayhawk Boulevard
Lawrence, KS 66045, USA
T 785 864-4645, F 785 864-5335

 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/seek-taxon/attachments/20060613/22144208/attachment.htm