Biology Reference
In-Depth Information
In other words, ontologies are not static lists, but rather dynamic struc-
tures that evolve with scientifi c ideas. Building ontologies, Smith con-
cludes, should be like building scientifi c theories. The vision for biologi-
cal ontologies is that they will enforce agreement about what objects
exist and the relationships between them—objects in the biological
world will have to be fi tted into the structure of the ontology.
Some social scientists have taken for granted the important role that
ontologies should play in encouraging collaboration and data sharing.
Sabina Leonelli argues that “bio-ontology consortia function as a much-
needed interface between bottom-up regulations arising from scientifi c
practice, and top-down regulations produced by governmental agencies.
They achieve this by focusing on practical problems encountered by
researchers who use bioinformatic tools such as databases.”
31
She sug-
gests that the centralization of power in consortia like the GO Consor-
tium actually promotes epistemic pluralism in biological research.
My ethnographic experience, however, suggests that ontologies in
general, and the GO in particular, are not universally accepted tools
among working biologists. This is best appreciated by noting that the
GO has several competitors; although the GO is the most widely used
gene ontology, biologists often run their data against multiple gene
ontologies before being satisfi ed that the result is plausible.
32
At a lab
meeting, one principal investigator warned his students and collabora-
tors: “GO is useless, I always ignore it . . . it's way too general.” Express-
ing her distrust of the GO, another bioinformatician reminded me that
ultimately the system was maintained by “a bunch of postdocs” who
just sat in a room reading scientifi c papers and deciding what terms to
include. When I raised the subject of ontologies among biologists, they
often reminded me of the quip, sometimes attributed to Ashburner, that
“biologists would rather share their toothbrush than share a gene name.”
Such reactions suggest a profound discomfort among biologists with
the centralization of responsibility and the structuring of knowledge that
ontologies impose. Biologists often describe their discipline as “messy”
compared with sciences like physics or chemistry; what is interesting
is found in the exceptions rather than the rules, and careers have been
built on the ability to navigate this uniqueness. This view suggests why
the kind of standardization that ontologies offer is not always well re-
ceived: biologists consider the freedom and fl exibility of their categories
and their language to be an advantage for investigating and describing
biological systems. Conforming to standards, even if they are collabora-
tively developed, and speaking in controlled vocabularies may not be in
a biologist's self-interest, at least not in the short term.
