Why are ontologies mostly used in biomedical context?

[phiofx]

Its a general question, not specific to owlready but maybe this community has some insights. In principle ontologies (owl and related concepts) could be applicable and used to good effect in many domains. But that is generally not the case and there is criticism, e.g., that the concepts and tools are cumbersome or abstract, or one can do things with more conventional databases and schemas etc. On the other hand ontologies are particularly well "productionized" in biomedical fields. Is there something specific about biomedical data that makes the technology more powerful? Is it something about a more academic rather than commercial culture of that sector (e.g. sharing ontologies). Is it some standardisation initiative that succeeded? Or maybe just accidental?

[H_Borgelt]

Hello phiofx,

I am only speaking from my own persective and do not have a proof for my thesis. However, as i am working in the a-typicall field of catalysis, i have come across several challenges which were hard for me to solve and might not be a problem in the biology domain. Some of these were limitation such as the lack of mathematics (summation and substraction of two values) and comparing data inside an ontology. For me this is pretty much essential as a catalyst is only an abstract concept which is highly conditional. I think a lot of other domains also strugle at simplifying relations in such a way. The domain of biology can often simplify relations in a binary fashion (either a gene is affecting an illness or not; "True/False") and thus do not face these kind of problems often.

As for the arguments around the more academic or commercial aspects, i think that their are a lot of commercial ontologie solutions in the biology domain, which might even be more succesful than some of the open source ontologies. However it might be difficult to stumble upon them as they are typically not listed in tools such as the ontology-lookup-service. In the field of engineering for example there are some possibly sophisticated ontology hiden as iso-standards, to which no service really points.

best regards,
Hendrik

[phiofx]

yes, what you say makes a lot of sense. I some sense owlready is increasing our expectations. Being python based it brings ontologies much closer to data science oriented workflows but depending on the domain there might be various challenges to overcome.