Biology Reference
In-Depth Information
a tree hierarchy, but where each term can have multiple parents. For instance,
if a particular gene was involved in making hexose, its GO terms would
include “hexose biosynthetic process,” “hexose metabolic process” (since any
biosynthetic process is [less specifi cally] a metabolic one), and “monosaccha-
ride biosynthetic process” (since hexose is [less specifi cally] a monosaccharide).
Terms can be semantically linked to one another through relationships such as
“is_a” or “regulates_positively.” GO conforms to the specifi cations of a Web
Ontology Language (OWL) since its designers expect biological databases to
be searched over the web (see the Conclusion for more details on the relation-
ship between bio-ontologies and web ontologies).
21. Gene Ontology Consortium, “Gene Ontology,” 25.
22. “The Consortium emphasized that GO was not a dictated standard.
Rather, it was envisioned that groups would join the project because they
understood its value and believed it would be in their interest to commit to the
ontology. Participation through submission of terminology for new areas and
challenging current representation of functionality was encouraged.” Bada,
“Short Study.”
23. Interview with Midori Harris, November 26, 2008, Hinxton, UK. In
fact, there are four full-time editors and roughly forty further contributing
researchers who have permission to write to (that is, edit) the GO Concurrent
Versioning System.
24. OBO Foundry, “Open Biological and Biomedical Ontologies.”
25. One such ontology is called the Basic Formal Ontology.
26. Smith, “What Is an Ontology?”
27. See http://www.fl ickr.com/groups/folksonomy/. See also Ledford, “Mo-
lecular Biology Gets Wikifi ed.”
28. Smith, “What Is an Ontology?”
29. Smith, “What Is an Ontology?”
30. Smith, “Ontology as the Core Discipline.”
31. Leonelli, “Centralizing Labels.”
32. For instance, the DAVID (Database for Annotation, Visualization and
Integrated Discovery) web resource allows users to combine ontology terms
and other identifi ers from many different sources. See http://david.abcc.ncifcrf
.gov/ and Sherman et al., “DAVID Knowledgebase.” The plausibility of GO
becomes particularly important in high-throughput experiments in which tens
or even hundreds of genes may be simultaneously up- or down-regulated by
some exogenous conditioning (a drug, for example). The lab biologist then
wishes to fi nd out whether these genes have anything in common. This can be
done by testing for “enrichment” of annotation terms among the set of genes.
For example, if 50 of the 60 up-regulated genes were annotated with the GO
term “cell-cycle,” the biologist might be able to conclude that the drug has an
effect on the cell cycle.
33. On the Semantic Web and data sharing in biology, see Ure et al.,
“Aligning Technical and Human Infrastructures.”
34. Field et al., “Towards a Richer Description.”
35. Again, how the machines are physically linked together in real space
was of usually of little importance (unless something went wrong). Machines
