Biology Reference
In-Depth Information
and all they're interested in is computation and they have no
interest in the underlying biology. And so they're busy doing in-
teresting computational things that are biologically completely
meaningless.
7
This comment, made by the Nobel laureate Richard J. Roberts, suggests
that the hostility toward computational biology is rooted in a sociologi-
cal division between individuals with different approaches to biological
problems. By name and defi nition, bioinformatics is a hybrid discipline,
built from biology and informatics. Indeed, there are two quite distinct
groups of people who inhabit the fi eld. The fi rst consists of individuals
with training in the life sciences who have turned to computation in
order to deal with larger amounts of data, fi nd more powerful ways
to analyze their data, or build computational models of the biological
systems with which they work. The second consists of individuals who
have been trained in computer science, mathematics, or engineering (or
in computational areas of physics) and who have turned to biology to
fi nd new kinds of problems to solve and new applications for their pro-
gramming and algorithm-building skills.
8
As one might predict from
Roberts's comments, these groups have different interests, different ap-
proaches to problems, and different attitudes about what constitutes
good biological work. They may occupy different academic depart-
ments and publish in different journals.
9
In fact, almost everyone I talked to in the i elds of bioinformatics
and computational biology complained about problems of communica-
tion across the computer science-biology divide. When I asked what
challenges awaited bioinformatics in the next few years, hardly anyone
failed to mention the fact that more people needed to be adequately
trained in both biology and computer science. Much of what was char-
acterized as “bad” or “useless” work in the fi eld came either from com-
puter scientists misunderstanding what problems were important to
biologists or from biologists failing to realize that the computational
problems with which they were engaged had been solved by computer
scientists decades ago.
Christopher Burge, in whose lab I worked, was unusually respected
in the fi eld because his background in mathematics and computer sci-
ence was complemented by an encyclopedic knowledge of biology; as
such, he was able to run both a computational lab and a wet lab side by
side. But even so, Burge had to pay constant attention to the disjunctions
within the fi eld: he had to ensure that his lab was balanced between
