Designing Life’s Layered Circuits: Tools of the Trade
Synthetic biology moves beyond the whiteboard.
In synthetic biology labs around the world, brainstorming has often begun at the same place: in front of a whiteboard. Marker in hand, researchers jot down the parts needed to form a new circuit, draw lines and arrows to show how they interact, and scrawl notes about how to assemble the parts into an appropriate whole.
“It’s usually based on intuition, and what we know has worked in the past,” says Timothy Lu, MD, PhD, who heads up the Synthetic Biology Group at the Massachusetts Institute of Technology.
The whiteboard has been used to design many novel genetic programs—whether aimed at turning bacteria into biosensors or forming networks of enzymes to churn out a particular product. But the way of the whiteboard might be fading. As circuits become more and more complex, and researchers move toward the design of larger networks and whole-cell programs, it’s becoming harder to manage all the required parts for a new project in hand-written dry erase.
“When I was looking at a simple circuit with two inputs, I could by hand iterate through all the possible states of the system,” Lu says. “Now, I’m interested in things with six or eight inputs, and intuition starts to fail.”
Costas D. Maranas, PhD, professor of chemical engineering at Penn State University, concurs. While synthetic circuits of a decade ago had a single switch and just a few inputs to alter genes, Maranas is trying to reconstruct and regulate the entire repertoire of pathways involved in a microbe’s metabolism.
And it’s not just the number of switches that adds complexity. Adding new enzyme activity into a bacterium is more complicated than just adding the enzyme. Take nitrogenase, for example, which Maranas and collaborators at Washington University would like to be able to control within a cyanobacterium. Getting the right levels of nitrogenase activity, he adds, doesn’t just mean having the right levels of gene and protein expression, but also accurately reproducing the light to dark transitions and providing sufficient energy in the form of ATP to power the nitrogenase.
To help manage this complexity, researchers are developing, refining, and applying computerized design programs that track the parts involved in their systems and pinpoint the best method to assemble a new circuit. There’s not yet one program that fulfils the dream of “plug and play” biology—where a few simple clicks choose the parts for the essential biological circuits and, voilà, synthetic life! But several programs are emerging as crucial to the field.
Inspired by Engineering
In the mid-2000s, Jean Peccoud, PhD, a computational synthetic biology researcher at Virginia Tech’s Bioinformatics Institute, was working on recreating the genetic networks that control cell division in yeast. Like other synthetic biologists, Peccoud viewed the components of the network—genes, promoters, ribosome binding sites, and terminators, to name a few—as discrete parts, with defined functions, that could be shuffled around between networks. But he realized that no software existed that could track which parts worked together, guide how the parts could be plugged into genetic circuits, and model how a proposed circuit would function.
“It seemed reasonable to assume that synthetic biology would need some computer-aided-design tools just like any other engineering discipline,” says Peccoud. CAD programs are heavily relied on by electrical and mechanical engineers, for example, to design electrical circuits or structures on the computer before they’re created and tested.
So his lab began developing such a program for biology. The result: GenoCAD, an open-source, synthetic biology CAD software. GenoCAD manages lists of genetic parts and gives users an interface where they can set design rules, apply them to their system and then assemble genetic parts into plasmids. It also includes a simulation engine to test new circuits.
In the December 2013 issue of ACS Synthetic Biology, Peccoud and two collaborators describe using GenoCAD to create a set of grammatical rules for building novel synthetic transcription factors from seven different types of parts. The program was able to generate eight possible designs that met all the rules governing what parts were required and what order they should fall in. The rules, which were derived from experimental information, can be revised and updated over time. As new synthetic circuits are tested in living cells, their success or failure can help guide the design of future circuits.
A Growing Toolbox
In addition to GenoCAD, there are a rapidly growing number of synthetic biology tools, Peccoud says. He adds that his ultimate goal isn’t for GenoCAD to beat out other tools. “I don’t think it should be a goal to converge to one tool,” he says. “Our field is so new that it is necessary that people explore different avenues.”
When designing DNA to characterize new promoters, George McArthur, PhD, a chemical engineer at Virginia Commonwealth University, turns to a different software program for nearly every step of the process. Aside from GenoCAD, he uses a ribosome binding site (RBS) calculator that develops an RBS of whatever binding rate he needs; a tool that produces inert spacer sequences; and the automated DNA assembly program J5 that gives him a list of primers for use in assembling the sequences he designs in GenoCAD.
“As a user, I’d love to have everything in one place,” McArthur says. “And already it’s great that a lot of these tools adhere to the same file standards. I think that eventually we’ll have different collections of software that aggregate together.”
One effort to encourage the consolidation of tools—or at least the development of a start-to-finish synthetic biology design protocol—is DARPA’s “Living Foundries: 1000 Molecules” program. Approximately $110 million in grants will be doled out by the end of 2014 to scientists who aim to build infrastructures for engineering biological molecules. The proof of principle for any infrastructure will be the design and production of 1,000 new molecules. But following through to such an outcome will likely require a strong set of software tools and provide an example for the rest of the field to follow.
Predictive Power Still to Come
Peccoud admits that the weakest part of GenoCAD is the newest addition to the program—the simulation engine. “Being able to run simulations of the behavior of a synthetic genetic system before making it is the holy grail of synthetic biology,” he says. “The science is not there yet but it is our hope that a tool like GenoCAD can help support the research necessary to understand gene expression better.”
Lu says that getting more accurate simulations of biological circuits will require more data on how different organisms interact differently with the various parts that make up circuits.
“In other engineering disciplines, the manufacturer of a system will give you parameters that define that particular system,” Lu says. Those parameters can be entered into CAD software to make your computer models accurate. “But in biology these days, no one has defined what, say, the E. coli parameter set is,” he points out. Even if a circuit is completely worked out in one strain of E. coli, he says, moving it to a new strain can drastically change how it functions.
Recently, Stanford scientists created a whole-cell computational model of the circuits within Mycoplasma genitalium, a human pathogen. Lu has collaborated with the Stanford team to start putting synthetic circuits into the organism. With the whole-cell model at his disposal, he hopes to start predicting how new circuits will work in the organism. But even that has been slow going, he says, and it’s just one bacterium.
Even without full predictive power though, programs like GenoCAD are pushing the boundaries of synthetic biology, offering a more modern “whiteboard” to sketch out complex circuit designs and organize growing libraries of biological parts. By giving users a place to organize the grammatical rules that govern their design process, and the parts that they want to use, it makes the design step of the standard engineering “design, build, test” cycle that much easier.