MIT Scientists Ask Biologists to Imitate Engineers

New projects try to apply interoperability and failure analysis to DNA experiments.

Biotechnology could enjoy more of the efficiency and innovation of high technology, say a group of experts working at the intersection of computation and biology. But first, researchers have to think less like scientists and more like engineers.

By inserting new collections of genes into bacteria, teams of "biohackers" can make microorganisms do nifty tricks: reproduce black and white photographs, form a ring oscillator, work as an etch-a-sketch under a ultraviolet pen, even detect whether coffee is decaf, regular or espresso.

But biologists cant take a trick invented by one lab and combine it with tricks invented by someone else.

"Its like an iPod that can only connect to your computer and your stereo," said MIT professor Drew Endy, in Cambridge, Mass. If every iPod had to be made by trail and error to fit one particular machine, he asked, how expensive would iPods be, and how many people would bother to own one?

In 2003, Endy and others set up the Standard Registry of Biological Parts. The idea was to get biologists to write interoperable plug-ins in DNA code. These plug-ins are described in an online database and stored in a freezer at MIT.

All researchers have to do is follow a set of assembly rules when designing their genetic gizmos, essentially making sure that the wiring required for one wont short-circuit another. That approach is standard operating procedure for many computer engineers.

But few biologists are putting their inventions into the registry. They dont want to design their genetic gizmos following assembly rules that would keep one gizmo from interfering with another. Instead, they continue to build what they need from scratch, making dozens and dozens of sequences until getting something that works, often only within the scope of their own project.

Most scientists have an "unengineering" mind-set, according to Tom Knight, a senior scientist at MIT who designed computer systems before turning to the intersection of computation and biology. "Scientists are like artists. They want to be able to do whatever they want" in whatever biological systems they are most comfortable in, he said.

/zimages/5/28571.gifClick here to read about how a supercomputer is serving as a weapon in the fight against AIDS.

Scientists embrace complexity, he said, while engineers shun it; part of the engineering mind-set is a willingness to follow standards. Standards might make an individual project less efficient, he said, but following them can also create tools that work 99 percent of the time.

Endy, trained as an engineer, rebels against the kind of inefficiency that biologists find an intricate part of science. Take the task of attaching two pieces of DNA to each other, an essential task for basic biological research and for making most protein drugs. "Its a 30-year-old technology," Endy said, but people routinely accept that its going to fail. "If it works half the time, youre a rock star."

Now Endy has hit on a new solution, The Office of Biological Dis-Enchantment, which Endy announced to laughter to Berkeley. Instead of asking biologists to design genetic gizmos to spec and contribute them to a registry, he said he wants to collect descriptions of failed experiments, put them in a database and mine for fundamental sources of error.

"Biological systems are not optimized to be easy to interact with," Endy said. Analyzing failure, an engineering approach, can help fix that problem, he said. Whether it can really bridge the gap between biologists and engineers remains to be seen.

/zimages/5/28571.gifCheck out eWEEK.coms for the latest news, views and analysis of technologys impact on health care.