Biology Reference
In-Depth Information
of the columns within them, meaning that a command or query would
still work even if data were reorganized; as can be seen in the library
example, the rows and the columns could be rearranged without affect-
ing the outcome of a query.
The relational model had two advantages over its “network” rivals.
First, it did not require relationships between data to be specifi ed dur-
ing the design of the database; second, the abstraction of the structure
from the physical storage of the data greatly simplifi ed the language
that could be used to manipulate the database. “Because the relational
model shifted the responsibility of specifying relationships between
tables from the person designing them to the person querying them,”
Haigh argues, “it permitted tables to be joined in different ways for dif-
ferent purposes.”
9
Relational databases present an open-ended, fl exible,
and adaptable means to store large amounts of data. Despite IBM's ini-
tial support for the “network” model, the development of Codd's ideas
through the 1970s led to the development of SQL (Structured Query
Language) and the commercialization of the relational model through
fi rms such as Oracle and Sybase.
Even from this brief history, it is clear that different types of database
structures are appropriate for different types of data and for different
types of uses. Moreover, this history suggests that databases act as more
or less rigid structures for containing information—that the proximity
and accessibility of particular kinds of data are determined by the form
of the database itself.
Dayhoff and a New Kind of Biology
The fi rst biological databases—that is, the fi rst groupings of biological
information ordered on a computer—were produced by Margaret Oak-
ley Dayhoff. Dayhoff, born in 1925, was trained in quantum chemistry
under George E. Kimball at Columbia University, receiving her PhD
in 1948. Her thesis work involved calculating the molecular resonance
energies of several polycyclic organic molecules—a computationally in-
tensive problem that involved fi nding the principal eigenvalues of large
matrices.
10
In approaching this problem, Dayhoff devised a way to use
punched-card business machines for the calculations. After her gradu-
ate studies, Dayhoff pursued her research at the Rockefeller Institute
(1948-1951) and at the University of Maryland (1951-1959). In 1960,
she joined Robert Ledley at the National Biomedical Research Founda-
tion (NBRF, based at Georgetown University Medical Center, where she
also became a professor of physiology and biophysics), and it was here
