Big things from Little b

The Little b language developed at Harvard Medical School is a Lisp derivative tailored for biological work. It maintains a library of "assertions" about biological processes and uses an expert system kernel called Lisa to tell the language what to do in various situations. It might recognize, without the user or programmer telling it, that because Enzyme E is in proximity to Chemical A in some experimental data, it is likely to produce Chemical B. It might then go on to deduce additional outcomes from the presence of Chemical B, says Jeremy Gunawardena, a mathematician and the director of the Virtual Cell Program at the school.

In addition to these assertions, or rules, Little b uses a database of reusable templates, a little like subroutines, that relieve the researcher and the programmer of the need to specify common biological scenarios each time they are encountered. Says Gunawardena, "In the past, we have not been able to abstract those generic components, so we have to keep re-describing the wheel every time we see it."

Aneil Mallavarapu, the chief developer of the language and director of the Little b program at Harvard, has degrees in biochemistry and cell biology but now considers himself a computer scientist. "I learned that independently," he says. "My father gave me a computer when I was very young and the games were terrible, so I started programming."

As a graduate student in biology, he was unhappy with available research tools, just as he had been with computer games earlier. "My understanding of how software engineers handled complexity really made me start thinking about how biologists might handle it," he says. He says he designed Little b for cell biology but with the idea in mind that its basic concepts -- its core language idioms, symbolic math and object-oriented definitions and reasoning -- might also be applied to complex disciplines such as economics or ecosystems.

Mallavarapu says computer science may be even more important to biology than mathematics. "Biologists are expected to understand sines and cosines and all kinds of weird trigonometric functions as a standard part of the canon," he says. "But we don't have that attitude toward computer science, which is arguably much more fundamental than a lot of the mathematics that we get taught."

Copyright © 2008 IDG Communications, Inc.

Shop Tech Products at Amazon