 |
 |
 |
The bio-zen Ontology Framework
The development of the bio-zen Conventional databases (e.g. relational databases or XML databases) do a poor job of representing biological reality. Researchers that want to publish or search for information do not only have to know about the biological structures they are investigating, they also have to deal with the structures of the database tables, file formats or XML documents – all of which are in most cases only remotely similar to the mental representation we have in mind when thinking about biological facts. Furthermore, most databases in use nowadays were designed with only a small and limited field of investigation in mind, and so we are now confronted with a convolute of small databases that can only be made to work together through a lot of additional work. We also see that systems biology with its focus on the simulation of complex biological systems has an ever growing impact on classical molecular biology. However, the realm of qualitative information that is represented in texts and databases like Uniprot or BIND is completely disconnected from the world of simulation and modelling, which is currently represented through languages like SBML, CellML or NEURON models. To realize the promises of systems biology, the division between these two worlds has to be bridged. The bio-zen framework uses Semantic Web technologies to overcome the limitations of current information systems in the life sciences. The descriptions of biological reality in the bio-zen framework are very similar to the cognitive models researchers have about their subjects of investigation, making the work with information systems more intuitive for the individual scientist. bio-zen is exceptionally flexible and extensible, making it easy to represent information from a wide variety of fields in a common framework. It also allows for a seamless integration of mathematical descriptions and simulation parameters into qualitative information, enabling a quick transition from data and information to model simulations and back. If successful, the Semantic Web and the bio-zen framework could mark the beginning of a whole new paradigm in the organisation and dissemination of information in the life sciences. The following Semantic Web ontologies are incorporated in bio-zen (all ontologies are mapped to DOLCE):
Basic design principles of the bio-zen framework: • The ontology is built upon existing foundational ontologies ( • Over-specialisation is avoided where possible. Generic properties like ‘part of’, ‘constituent of’, ‘broader concept than’. ‘caused by’ etc. are the backbone of the whole ontology. The class hierarchy is kept simple and lean. • Dublin Core and other existing metadata standards are used. • The ontology integrates two different approaches of information representation in a common framework: ‘realist’ ontological descriptions and ‘conceptualist’ taxonomies and concept hierarchies. - Making a clear distinction between both approaches reduces the susceptibility to inconsistencies. - Unifying both approaches in one common framework makes it possible to combine the specific advantages of each approach in the best way possible. The consistency of the realist approach is complemented with the flexibility of the conceptualist approach. • Users of the ontology only need to make OWL individuals to represent information. The definition of new classes is not necessary, which helps to avoid many of the associated problems. • The ontology is focused on the description of spatio-temporal particulars (concrete biological things and models). This focus limits the problems that arise when users of the ontology can define universals. • The ontology allows the seamless integration of mathematical models (similar to and compatible with SBML 2) into qualitative information. • Molecular interactions are modelled as stochastic processes involving populations of molecules, not as singular events that involve singular molecules. This approach is much closer to biological reality in most occasions, and avoids some grave consistency problems associated with the other approach. • Strict use of SI units for physical measurements. • The ontology includes simple means to 'fuzzify' information (i.e. basic support for fuzzy logic - like constructs).
The first version of the bio-zen manual is ready for download (PDF, 500 kb) |
|
ontology
|
