Wednesday, January 25, 2006
Generative Methodology Glass Bead Games
Scholarly exchange within the BioPAX working group
> For example, given that in your suggestion you used a transitive
> property, I might guess that given a gene, you might like to find
> all the gene products and that you might do that by asking
for the
> encodes property values of the gene, expecting that you retrieve
> all the transcripts (direct encodes) and proteins (transitive
> encodes) that correspond to them.
Yes exactly, it merely should serve as a simply way of associating genes and their products. For example, Pubmed databases offer the "Links" - link for each entry, where you can navigate between gene, transcript and protein; and many other databases have similar features. It should not be an attempt to model more complex biological processes associated with transcription/translation, like the extensions in the 'gene regulation' proposal do. However, the more expressive concepts in the proposal might be desirable for many users. The question is: do the more expressive concepts in the gene regulation proposal allow the user to make the 'quick and dirty' associations between biopolymers, or does it force him/her to make things more complex than they really are?
Comments by Paul Prueitt
As we all realize, the term
"expressive" is overloaded .... one interpretation having to do
with "logical expressiveness" (where the definition is not uniform in
the literature and where expressiveness implies that formal first order logics
"ought" to be able to model any natural phenomenon)
the other interpretation has to do with biological expressiveness.
Biological expressiveness is not sameAs logical expressiveness..... (conjecture)
This conjecture has not been disproved, and the assumption of the negation of this conjecture is problematic for Protege-OWL.
The use of "link" as a property is an excellent one and leads directly into the n-ary (without logical apparatus) approach to modeling "reality" using computer encodable data structures..
See continuing discussion.. à [143]