Drug a role or entity?

[Egon Willighagen]

Hi all,

I noticed that a Drug is currently a specialized entity? But shouldn't
this be a role instead? The description does refer to a process as a
condition:

"A drug is a biochemical entity that *regulates* a biological
process." (emphasis mine)

See: http://semanticscience.org/resource/SIO_010038

What am I missing here? Or have we simply be picky on SIO, and use
some of it, but not all in CHEMINF?

Egon

--
Dr E.L. Willighagen
Postdoctoral Researcher
Institutet för miljömedicin
Karolinska Institutet (http://ki.se/imm)
Homepage: http://egonw.github.com/
LinkedIn: http://se.linkedin.com/in/egonw
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers

[Michel Dumontier]

Hi Egon,

  (adding this to SIO's mailing list)

 admittedly, while I'm working on a theory of qualities and realizables (dispositions, capabilities, functions, roles), very little of this have concretely translated into SIO proper. However, the main idea is that role-holding entities such as 'drug' are objects that play a particular role R in some set of circumstances T. We currently recognize three kinds of roles - abstract (e.g. comparative, positional), processual (realized in only processes e.g. drug) and social roles (meeting certain social/legal requirements and/or conferred on by society e.g. doctor).  In the case of processual roles, we want to provide the basis of the role in terms of anatomical parts and their qualities and dispositions, and also describe the conditions, triggers and processes in which these are realized. 

  

  As far as CHEMINF is concerned, none of this applies, as the ontology is wholly focused on chemical descriptors/attributes for which their generation depends on algorithms and parameterized software.  

  

  Several OBO ontologies (e.g. ChEBI) do contain molecular roles such as drug:

CHEBI:23888

Any substance which when absorbed into a living organism may modify one or more of its functions. The term is generally accepted for a substance taken for a therapeutic purpose, but is also commonly used for abused substances.

http://bioportal.bioontology.org/ontologies/1007/?p=terms&conceptid=CHEBI:23888

 You can see that the definition specifies that instances of this class are substances, but we'd expect roles and substances to be disjoint. I would prefer to distinguish the role of 'drug', which 'drug role'.  I'll be looking at these and others to inspire a more complete role hierarchy for SIO, which also attempts to describe the basis and conditions under which these roles are held and realized.

m.

--
Michel Dumontier
Associate Professor of Bioinformatics
Carleton University
http://dumontierlab.com