Perhaps because I missed the first meeting and some of the conversations prior to that I struggled somewhat with where we were trying to operate in the generality->specifics spectrum. As such, maybe some of what I say below has already been discussed and rejected previously, if so, please ignore!
While I can see certain merits in choosing one low level specific example (physical instances and locations) and trying to work out the details of that as a prototype, I think there are also associated pitfalls.
Firstly, details are details, they are the most likely thing to vary between groups and between semantic ontologies. I'm not convinced that its easier to make our ontology flexible and open to external interfacing by working on the low level details rather than the more general concepts.
Secondly, the low level items are also more likely to change in time and perhaps become entirely obsolete. One (facetious) example would be the physical location of DNA. Once Drew finishes his Stinkjet and we can all print our sequences at will, then the long term storage of certain sequences seems less important.
I think putting together this ontology is a great and forward-looking thing to do. However, I think that pinning down the practical details of something that is inherently of a more long term benefit is very challenging.