Ontological foundations for biology knowledge models.

This paper analyzes the ontological requirements for representing biology knowledge, and identifies several areas where current knowledge representation (KR) paradigms need to be extended. We focus on the representation of experimental materials and methods, and the reasoning task of intelligent information retrieval; however, the ontological issues we raise apply to biology (and experimental sciences) in general. We have identified two important concept types in molecular biology that cause problems for standard knowledge models: 1) complex substances such as mixtures and nucleic acid sequences; 2) transformations (such as biochemical reactions) that convert one substances into another. We describe these problems, propose solutions for some of them, and given examples of the need for such knowledge representations in intelligent information retrieval.