Building Models

To make sense of such massive amounts of data, researchers create models. But there are three different ways of doing that.

The first might be called a "bottom-up" approach. Scientists first gather as much information as they can about all the components of a system, then try to combine that information into a complete picture of the system. Pharmacy researcher Tanja Kortemme, for instance, is making computational models of how proteins interact with each other. Kortemme's models enable her to analyze how the position of each atom in the protein allows it to hold or "dock" with another molecule. When the structure of all the proteins in a network is known, it should ultimately be possible to use Kortemme's method to model how the network works, by knowing how the pieces interact.

"Once you know the structure of proteins, you can predict that two different proteins would interact theoretically, even if you haven't yet found evidence of that in the laboratory," Kortemme says.

Kortemme's research is an example of how one method or technology can be globally applied to sweep up information about all active genes and proteins in a cell. Another example is a computer model that Kortemme's colleague in the School of Pharmacy, Andrej Sali, has created to predict how all proteins in yeast cells are folded from chains of amino acids. He is now modeling the more than half a million proteins that are identifiably related to one structure in a cell.

When finished, such a detailed model should be capable of describing exactly how the system works under any circumstance. "What we would like to do is be able to describe how a single gene behaves in the same detail that physicists can now describe exactly how hydrogen atoms behave in a gas," Voigt says.

At the opposite end of the methodological spectrum from the "bottom up" approach is the "top down" approach. This method starts with a model of how the researcher thinks the system works and then compares the model with information from the real biological system.

This approach borrows from a field called "complex adaptive systems" and utilizes powerful computer theories that model everything from traffic jams to weather patterns. The top-down models are the basis of computer games like "The Sims," in which individual players in the simulation are given rules of behavior and then allowed to interact freely with each other. It is most useful in situations in which the system is so complex that it cannot be taken apart easily.

"The theory of complex systems says that the rules that parts of the system go by are more important than a description of the parts," says UCSF researcher Tony Hunt. The system is also able to model behaviors that cannot be foreseen by looking at the individual players, he adds.

Another way to put it is that the whole is greater than the sum of its parts. "You can't take a part out of the system to study it, because its behavior in the test tube is different from its behavior in the organism," Hunt says. Hunt uses what he calls a "middle-out" approach -- mixing computer simulations and data from physical experiments -- to study liver chemistry. He notes that the activity of a liver cell varies greatly depending on its interactions with the cells around it.

"Liver function can't be understood simply by studying isolated liver cells in the lab," Hunt says. "That would be like trying to understand the legal system by observing a lawyer in an empty room."

The solution to this quandary is to make a virtual model of cell behavior on the computer. Hypothetical rules of cell behavior are formulated from what we know about liver cells, and then this model is refined as it is tested against the reality of the system's behavior as seen in experiments, Hunt says.

The advantage of this approach, Hunt says, is that we don't have to study and define the behavior of each cell individually, which would be impossible anyway. "A truly complex system simply cannot be understood through a reductionist approach," he says. "A complete description of the system is simply too big for us to hold in our heads."

The Machinery of Life

No matter how the models are arrived at, such accurate working descriptions of biological systems will change the nature of biology, the scientists say. With an understanding of all the biochemical parts of the cell and how they work together, scientists will be able to treat the cell as a machine that can be taken apart and reassembled in new and interesting ways. In short, engineered.

The problem with nanotechnology, as Chris Voigt points out, is that the motors and gears that researchers currently etch from silicon don't really work very well yet. He suggests that we may already have those micromachines under our noses -- and in our noses.

"Bacteria have already evolved machinery that works quite well," Voigt says. Once you understand how the pieces of that machinery fit together and work, you can mix and match pieces to create bacterial robots that can perform complex tasks.

Voigt's lab is taking the lead in this effort. By exchanging genes between organisms, he can add or subtract the protein machinery that makes the cell run. He can also link these protein pieces to each other to give cells new capabilities. One example is that Voigt's lab is now looking at how to control bacteria with lasers. He hopes to do this by taking specific genes from one bacterium that changes its behavior due to the presence of specific wavelengths of light, then linking those genes with other genes he wants to control. By putting the package into another bacterium and shining laser light on it, he can turn on or off whole biochemical pathways. Through selective mutation of the genes involved, he can even choose the wavelength of light he wants the bacteria to be sensitive to, he says.

The potential of such "synthetic biology" is so extensive that it inspires ideas that might seem far-fetched were others not taking them seriously. Recently, Voigt was awarded a grant from the National Academy of Sciences to hold a conference on using engineered microbes to assist in exploring Mars.

The idea of using microbes on Mars came about because Voigt wanted to promote new thinking in the field. "I realized that I needed some really out-there application that could pull together people who could contribute, but who don't necessarily see themselves as part of the synthetic biology world," Voigt says.

Bacteria, Voigt points out, have the advantage of being able to reproduce and spread. As an example of how they might be used on Mars, Voigt mentions the possibility of creating microbes that can move toward a particular chemical source, like methane or water, and signal back information about what they sense. Or spacecraft might be coated in an engineered bacterial biofilm that senses when there is a crack in the satellite, signals the presence of the crack through bioluminescence, and excretes a compound to fix it.

Once scientists and engineers understand cell systems well enough and open their minds to the potential of engineered organisms, they will come up with thousands of valuable new jobs for microbes, Voigt says. "They might be called upon to do everything from cleaning up toxic chemicals to exploring other worlds -- the possibilities are endless," he says.

Page 1 - The Rise of Systems Biology