Biology Reference
In-Depth Information
that it “permits biologists to describe and organize their data in a man-
ner that closely resembles how they typically think about their informa-
tion.” Indeed, when using the genome browser, one has the impression
of turning a microscope on the genome—as you zoom in far enough,
you begin to see the tiny As, Gs, Ts, and Cs emerge in the tracks at
the bottom of the window. This illusion of direct representation—the
notion that we are really navigating around a one-dimensional space
provided by the genome—no doubt contributes to the browser's useful-
ness. But the browser is also a tool for making the genome visible and
legible: in the browser, the genome is totally visible and manipulable,
it is a territory that is on display for quantifi cation and analysis. In-
deed, biologists often talk about using the browser for “data mining”—
discovering regions of interest and extracting biological meaning and
knowledge from them. The “illusory” quality of the browser is what
gives it power—it allows the user to see hidden connections between
tracks; it acts as a way of making biological knowledge by drawing
objects into juxtaposition. 42
Of course, the user is not actually “seeing” the genome: a genome
is not even a string of As, Gs, Ts, and Cs, but rather a coiled-up set of
macromolecules with some repetitive structural elements. Genes, exons,
and binding sites are not labeled, not visible through any kind of micro-
scope. Rather, the genome browser image is the result of an elaborate set
of computations and manipulations of sequence data. Just as a political
map shows imaginary boundaries that correspond to no visible feature
on the ground, the browser allows biologists to inscribe categories and
order on the genome in order to make sense of it. A track showing the
positions of genes, for example, might refl ect the output of a program
like GenScan that uses a hidden Markov model to predict the positions
of genes within a genome. 43 Tracks showing information about con-
servation between species are computed from painstaking attempts to
align genomes with one another, processes that themselves require large
amounts of computational time and power. Even the very positional
information on which the genome sequence is based has been computa-
tionally reconstructed from millions of shotgun-sequenced pieces. The
biological object is made visible through a series of re-representations
and computations. The user, then, is not looking at the genome, but
rather seeing the results of a series of analyses of biological information
(both genetic and nongenetic). The picture the user sees is not a repre-
sentation, but a computation.
The problem of translation between biology and computing is a
problem of representation. It is in fi nding practical and computation-
Search WWH ::




Custom Search