Skip to main content


November 20, 2002

Media Contact:
   Denine Hagen, (858) 834-2920 or


Bernhard Palsson, Professor
UCSD Bioengineering
Jeff Hasty, Professor
UCSD Bioengineering

In a study published in the November 14 issue of Nature, Bioengineers at the University of California, San Diego (UCSD) Jacobs School of Engineering used their computer model of E-coli (patent pending) to accurately predict how the bacteria would evolve under specific conditions. The results may have applications for designing tailor-made biological materials for commercial uses or for predicting the evolution of drug-resistant bacteria.

“This is totally revolutionary—that you can actually predict the outcome of such a complicated and intricate process as adaptive evolution,” says Bernhard Palsson, UCSD Bioengineering Professor and study author. “One of the implications of this study is that we could possibly use such a system to predict the evolutionary stability of bacteria, and potentially predict the probability of a drug-resistant strain developing.”

The study also serves as an example of the power of systems biology, a hot emerging field dedicated to employing mathematics and computer simulation to understand how genes and proteins work together to control the function of cells. Nature dedicated its November 14 Insights issue to the topic, which included an overview article co-authored by UCSD Bioengineering Professor Jeff Hasty.

Palsson first created a computer model of E-coli in 2000, and since then has shown that the model accurately mimics the behavior of the bacteria 80% of the time. He says he came upon this latest breakthrough almost by happenstance when his laboratory experiment of E-coli growing on glycerol did not match the rate of growth predicted by the computer model. On a hunch, he guessed that this particular strain of E-coli had not been exposed to glycerol before, and that if he gave the bacteria time to evolve, it might reach an optimum growth rate. To test the theory, Palsson created a “survival of the fittest” experiment, in which bacteria that grew well in glycerol was allowed to survive while less fit versions died off. He allowed the bacteria to evolve, which took about 40 days through 700 generations. The growth rate of the surviving strain matched the optimal growth rate predicted by the computer model. With this success in hand, Palsson’s group replicated the in silico-to-laboratory experiment with a number of different substrate materials.

Although beating drug-resistant bacteria is a foreseeable use of the technology, Palsson says a more immediate application is for tailoring microbes such as E-coli to make chemicals used in the synthesis of drugs and other products such as detergents.

“This is could be a totally new technique for designing commodity microbes,” says Palsson. “We could design a strain in the computer by adding or removing genes and then calculating the optimal performance of that strain. Once we have a strain that performs to the characteristics we want, we could move on to the real organism, manipulate the genetic content, and then use the adaptive evolutionary process to implement the design.”

This discovery by Palsson follows on another study reported in the November issue of Genome Research, in which Palsson used his computer model of the red blood cell to relate specific genetic mutations to exact variations of hemolytic anemia. It is the first model-based system for predicting phenotype (function of the cell or organism) based on genotype (an individual’s DNA). Both studies illustrate how new knowledge can be gained by creating computer models of how cells function—so-called genetic circuits.

“Every cellular function is a system requiring the overlapping interaction of dozens of gene products, and the coordinated action of multiple gene products can be viewed as a network, or a ‘genetic circuit,’ ” says Palsson. “These genetic circuits represent cellular wiring diagrams. They are the collection of different gene products that together are required to execute a particular function such as metabolism.”

Over the past two decades, Palsson has been working at the enormous challenge of creating computer models of these biological functions. He employs a technique he calls constraints-based modeling—basically describing what a cell DOES NOT do in order to define what it can do through a process of elimination. To date, Palsson has created in silico models of metabolism for E-coli, the red blood cell, H. influenzae, H. pylori and yeast.

Palsson’s lab is one of the few in the country to build a complete network of the circuitry in given cells. Other researchers in the field are describing simple control modules within the network of a cell, as outlined by UCSD Bioengineering Professor Jeff Hasty in an Insights article in the November 14 issue of Nature.

“There are small mechanisms within the circuitry of cells which can have a major impact on function. Just as in an engineered electronic circuit, each module performs specific duties, such as switching a protein on or off, or generating oscillations in the amount of protein released based on the time of day,” says Hasty.

Researchers are beginning to build and test synthetic versions of these control modules. For example, Hasty has developed a model of a positive feedback loop, in which a gene produces a protein which in turn causes that gene to become more active. He says as scientists begin to synthesize these simple network modules in the context of mathematical models, it will set the stage for the controlling cellular function, which could have important applications in nanotechnology and gene and cell therapy. Hasty's long-term goal is to build synthetic genetic networks which could be inserted into a patient's cells to tightly regulate the expression of a desired protein, or even to cause an undesirable cell to self-destruct.



Print News Release  Email News Release

Search News


Subscribe to our Newsletter

RSS Feeds