ChemAxiom: An Ontology for Chemistry 3 – Choosing an Upper Ontology

I have to start this blogpost with a big mea culpa. In the mists of time, I proclaimed loudly and visibly on this blog, that I thought that Upper Ontologies were, well, a bit……ya’ know…..

I have realised, that this view is wrong and entirely misguided and that Upper Ontologies are needed for the construction of modular and integratable ontological systems. But before delving into this discussion, what are Upper Ontologies?

Upper Ontologies are concerned with describing modes of existence and being and within computer science they are used for preparing computable representations for the modes of existence of things, Wikipedia actually does a fairly good job of explaining further:

The aim is very broad semantic interoperability between a large number of ontologies accessible “under” this upper ontology. As the metaphor suggests, it is usually a hierarchy of entities and associated rules (both theorems and regulations) that attempts to describe those general entities that do not belong to a specific problem domain.

The seemingly conflicting use of metaphors implying a solid rigorous bottom-up “foundation” or a top-down imposition of somewhat arbitrary and possibly political decisions is no accident – the field is characterized by controversy, politics, competing approaches and academic rivalry.[citation needed]

Debates notwithstanding, it can be said that a very important branch of an upper ontology can be considered (as continuation and development of natural philosophy) to be the physical ontology.

Now the Wikipedia article goes on to point out several upper ontologies in common use in information science today. The most prominent are (a) the General Formal Ontology (GFO), (b) the Basic Formal Ontology (BFO), (c) the Suggested Upper Merged Ontology (SUMO) and (d) the DOLCE ontology.Again, the Wikipedia article on upper ontology does a fairly good job at summarizing the differences between and the peculiarities of these various upper ontologies. Now with so much on offer, the question remains, what the criteria are for choosing an upper ontology that would be appropriate for the tasks, which ChemAxiom is trying to accomplish.

First and foremost, any upper ontology used in the ChemAxiom project must have the sufficient scope to describe the phenomena associated with chemical objects, which I have laid out in a previous blog post: chemical objects have both synchronic and diachronic qualities, which, in turn, have very different associated ontologies. Secondly, any upper ontology which we use should facilitate the broadest possible uptake of ChemAxiom by the community, either because (a) related ontologies in other domains use the same upper ontology and the upper ontology has a good acceptance or (b) because relatively straightforward mapping between ontologies is possible (for example, the mapping between many BFO and SUMO concepts is relatively straightforward).

Taking the above criteria into account, we have chosen to use the Basic Formal Ontology for the purposes of developing ChemAxiom. The BFO has many things going for it. First of all, it has a notion of both syn- and diachronicity, each of which is codified up in a “sub-ontology” but integrated via the notion of an Entity:

bfo:Entity
      a       owl:Class ;
      rdfs:label "entity"^^xsd:string ;
      rdfs:subClassOf owl:Thing ;
      owl:unionOf (snap:Continuant span:Occurrent) .

Typical examples of continuants are molecules, chemical identifiers, substances etc, whereas occurents can be chemical reactions, measurements, interconversion phenomena etc… The BFO is therefore nicely set up for talking about these various qualities (in the non-BFO sense) of chemical objects and thus fulfills the requirement of sufficient scope outlined above. This is even further re-enforced by the BFO’s notion of “Role”. In the words of the BFO, a role is a

“realizable entity [snap:RealizableEntity] the manifestation of which brings about some result or end that is not essential to a continuant [snap:Continuant] in virtue of the kind of thing that it is but that can be served or participated in by that kind of continuant [snap:Continuant] in some kinds of natural, social or institutional contexts. […] Examples: the role of a person as a surgeon, the role of a chemical compound in an experiment, the role of a patient relative as defined by a hospital administrative form, the role of a woman as a legal mother in the context of system of laws, the role of a biological grandfather as legal guardian in the context of a system of laws, the role of ingested matter in digestion, the role of a student in a university.”

The tool of a “Role” is incredibly useful when trying to describe certain generic types of molecules or substances, such as acids, catalysts, nucleophiles, solvents, etc….A solvent, for example, can be modelled as a chemical substance which has a role “SolventRole” (which, in turn, is a subclass of role) The BFO is provides useful mechanisms for these things and biological roles of chemical compounds can potentially be dealt with in the same way.

Measurements, by comparison, are not continuants, but processes – occurrents. The same is true for interconverting isomers and other fluctional processes associated with molecules and substances, as well as phase changes, transitions etc…and the BFO provides a mechanism to deal with those phenomena.

Finally, the Basic Formal Ontology is the ontology adopted by the OBO family of ontologies and thus fulfills the criterion of wide acceptance within the community of practice or a related community – in this case bioscience and medicine. This should help to facilitate the integration of ChemAxiom with those ontologies, should this be a desireable thing for the community. I think, the blogposts until now set the scene nicely for some more detailed discussions of the individual ontologies, which will follow in subsequent posts either during or after the holidays. Until then, Happy Easter everyone. Comments and suggestions via the usual channels please…autogenerated links and tags – as always – by Zemanta.

Reblog this post [with Zemanta]

One Response to ChemAxiom: An Ontology for Chemistry 3 – Choosing an Upper Ontology

  1. Phil Lord says:

    It’s worth bearing in mind, thought, that BFO’s treatment of roles have some problems. The biggest
    one is that only independent continuants can bear a role. So, for example, data (which is a generically dependent continuant) cannot. So if you thought that some data was a result (a role) of an assay, this is in BFO speak not an accurate understanding of reality.

    Likewise, neither processes nor qualities cannot have qualities. So, if you want to talk about things like the rate of change of concentration of a molecule, again you have misunderstood reality. Concentration is a quality and cannot therefore have a quality itself. Likewise, chemical reactions which are processes can’t have a rate or otherwise as qualities.

    Of course, you can put these statements in outside of BFO. BFO doesn’t stop you (which is rather hard to ontologically anyway which tend to use open world semantics), but then this somewhat defeats the point of using BFO.

    Not that you shouldn’t use BFO, of course. All upper ontologies have their problems, but I think you will find these issues problematics.

    Good luck!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: