Biology Reference
In-Depth Information
data to a two-dimensional array provides both a powerful computa-
tional tool and a cartographic metaphor for relating to sets of data. In
a biology of massive data sets, high throughput, and statistics, it is the
visual that is coming to dominate the textual. Sequence is no longer just
script. Computer-based visualization has made (and is making) possible
the comprehension of vaster and more heterogeneous data sets. It is
therefore helping to justify the massive production of data that charac-
terizes contemporary bioinformatics.
The importance of computer graphics to biology is largely a recent
phenomenon. Before the 1980s, most computers were simply not pow-
erful enough to perform image processing. Even before computers could
be used as visualization tools, however, biologists deployed visual meta-
phors for solving data-intensive problems. The fi rst part of this chapter
describes how maps and matrices became ubiquitous features of data-
driven biology in the 1960s and 1970s. The increasing availability of
graphical workstations and personal computers in the 1980s allowed
biologists to develop more sophisticated representational tools. The sec-
ond part describes the development of AceDB for the representation of
the worm genome and its evolution into today's genome browsers. The
last part concentrates on how visualization tools such as genome brows-
ers and heat maps are actually used in the day-to-day work of biology.
These visualization tools are central to how biologists see, manipulate,
and know biological objects. The forms these images take are products
of the histories of computers, algorithms, and databases.
From MAC to Mac: How Computers Became Visualization Tools
In 1966, Cyrus Levinthal began to use computers to visualize pro-
tein molecules. This effort was part of the Advanced Project Research
Agency's Project MAC at the Massachusetts Institute of Technology.
5
Levinthal and his co-workers saw their work as a way to replace the
ungainly and bulky space-fi lling physical models that were used to
represent proteins. The aim was to allow viewing and manipulation of
large macromolecules without the hassle of actually having to build a
three-dimensional representation from wood, metal, or plastic. Levin-
thal attempted to achieve this by designing an elaborate set of systems
for interaction between the user and the model on the computer screen:
a “light-pen” and a trackball-like device allowed (to quote Levinthal's
paper) “direct communication with the display.”
6
As Eric Francoeur and Jérôme Segal have argued, these protein pic-
tures provided a middle way between physical models and purely nu-
