Things change—develop, mature morph—but not everything in the same way. Representing this knowledge in biomedical ontologies faces issues on three fronts: what the category of the participating objects are, which type of relations they involve, and where constraints should be added. Taking the transformation_of relation from the OBO Foundry Relation Ontology [1] as example, one can dress up this relation with various constraints that deal with the temporal aspects of the intention of that relation, but in such a way that it is still possible to represent it in an a-temporal ontology language. This can be done by using some basic notions of the OntoClean approach [2] to ontology engineering and generalising the notion of status classes from formal temporal conceptual data modelling into a novel status property. Criteria are identified, formulated in 17 additional constraints, and assessed on applicability for representing transformations more accurately, thereby enabling developers of bio(medical) ontologies to represent and relate entities more precisely, such as monocyte & macrophage and healthy & unhealthy organs. You can find the details in [3].
Rest me to add one comment for BFO-groupies: note that I am using OntoClean, not DOLCE, so please do not waste yours and my time with [a / another] rant that I am not allowed to use DOLCE for the realist-based RO, because I do not. OntoClean and DOLCE are orthogonal (hence, so are OntoClean and BFO & RO) and OntoClean is independent of the debate on realist-versus-other ontological commitment.
References
[1] Smith, B., et al.: Relations in biomedical ontologies. Genome Biology, 6 (2005) R46.
[2] Guarino, N., Welty, C.: An overview of OntoClean. In Staab, S., Studer, R., eds.: Handbook on ontologies. Springer Verlag (2004) 151-159
[3] C. Maria Keet. Constraints for representing transforming entities in bio-ontologies. KRDB Research Centre Technical Report KRDB09-2, Faculty of Computer Science, Free University of Bozen-Bolzano, Italy. April 22, 2009.
Keet blog 