next up previous
Next: Summary Up: A Survey of Current Previous: The Gene Ontology

The TAMBIS Ontology

TAMBIS (Transparent Access to Multiple Bioinformatics Information Sources) uses an ontology to enable biologists to ask questions over multiple external databases using a common query interface [1]. The TAMBIS ontology (TaO) [19] describes a wide range of bioinformatics tasks and resources, and has a central role within the TAMBIS system.

An interesting difference between the TaO and some of the other ontologies reviewed here, is that the TaO does not contain any instances. The TaO only contains knowledge about bioinformatics and molecular biology concepts and their relationships - the instances they represent still reside in the external databases. As concepts represent instances, a concept can act as a question. The concept Receptor Protein represents the instances of proteins with a receptor function and gathering these instances is answering that question.

The TaO is a dynamic ontology, in that it can grow without the need for either conceptualising or encoding new knowledge. In contrast, the other ontologies described here are static - developers must interveen and encode new conceptualisation to form new concepts. The TaO uses rules within the ontology to govern what concepts can be joined to another concept via relationships, to form new concepts. Thus the TaO places great emphasis on relations. A user can form a complex, multi-source query, using relationships, in the following manner. Starting with the concept Protein, the TaO is consulted as to which relationships can be used to join Protein to other concepts. Amongst many, the following two are offered: is homologous to Protein and hasAccessionNumber AccessionNumber. Initially, the original Protein is extended to give a new concept Protein isHomologous to Protein (The concept Protein Protein homologue); then the second `protein' is extended with hasAccessionNumber AccessionNumber. The resulting concept (`Protein homologue of Protein with Accession Number') describes proteins which are homologous to protein with a particular accession number. This concept can be used as a source independent query containing no information on how to answer such a query. The rest of the TAMBIS system takes this conceptual query and processes it to an executable program against the external sources [20].

The TaO is available in two forms - a small model that concentrates on proteins and a larger scale model that includes nucleic acids. The small TaO, with 250 concepts and 60 relationships, describes Proteins and enzymes, as well as their motifs, secondary and tertiary structure, functions and processes. There is also supporting material on subcellular structure and chemicals, including cofactors. Motifs extend to detail such as the principal modification sites; function and process to broad classifications such as Hormone and Receptor, and Apoptosis and Lactation; structure extends to detail such as gross architecture - for example, SevenPropellor. Important relationships include is component of, has name, has function and is homologous to, as well as many more. The larger model, with 1500 concepts, broadens these areas to include concepts pertinent to nucleic acid, its children and genes.


next up previous
Next: Summary Up: A Survey of Current Previous: The Gene Ontology
Robert Stevens 2001-07-19