Biology Reference
In-Depth Information
Ostell had begun to experiment with the MOLGEN system for ana-
lyzing his sequences. He found the tools it provided unsatisfactory for
his purposes. As a result, he began to write his own sequence analysis
software in FORTRAN, using a compiler that had apparently come with
the computer. In his dissertation, written some seven years later, Ostell
outlined two sorts of differences between the MOLGEN programs and
his own. First, the MOLGEN software was designed to run on a main-
frame, “supported by substantial government grants.” By contrast, the
system that Ostell was using was “mainly the province of computer
buffs and cost about $10,000. . . . It was a radical idea that comparable
performance could be attained from a (relatively) inexpensive desktop
computer.”
62
Second, Ostell's programs were user-friendly:
The software attempted to converse with the scientist in an
immediately understandable way. Instead of questions like
“>MAXDIST?_,” such as one would encounter on the Molgen
system, this package would ask things like “What is the maxi-
mum distance to analyze (in base pairs)?” The other aspect of
“doing biology” was the way the analyses were done. For exam-
ple, the Molgen software would give the positions of restriction
enzyme recognition sites in a sequence. But why would a biolo-
gist want to do a restriction search in the fi rst place? Probably
to plan a real experiment. So my package would give the cut site
for the enzyme, not the recognition site. . . . I feel mine provided
more immediately useful information to the scientist.
63
Ostell's colleagues, fi rst in his own lab, then all over the Harvard Bio-
Labs, soon began asking to use his programs. When he published a
description of the programs in
Nucleic Acids Research
in 1982, offer-
ing free copies to anyone who wanted it, he was overwhelmed with
requests.
64
Ostell's programs constituted one of the most complete
software packages available for molecular biology and the only one
that would function on a microcomputer. In addition to making his
programs microcomputer-friendly, Ostell made sure that they could be
compiled and used on multiple platforms. Roger Staden's similar pack-
age suffered from the fact that it used unusual FORTRAN commands
and made occasional PDP-11 system calls (that is, it was designed for
a mainframe).
65
Over the next few years, Ostell occupied himself with
making the package available to as wide a range of collaborators as pos-
sible, adapting it for different systems and adding additional features.
