VOLUME 12, NUMBER 14 • JANUARY 26, 2001

Physics and biology intertwine at Rockefeller

Rockefeller University’s Center for Studies in Physics and Biology began approximately seven years ago when some of the faculty and administration at the university began to sense the ways that biology and physics might be poised to enter a new dialogue. In the short time since its inception, the center has demonstrated prescience not uncommon to the university. For quite a while, Rockefeller was the only scientific institution in the country supporting such an alliance between physicists and biologists and even today, as other universities are scrambling to create centers for physics and biology, Rockefeller’s remains the most powerful and productive.

Though relatively small, the center has a strong identity. Professors Albert Libchaber and Eric Siggia and Associate Professor Marcelo Magnasco, for example, have carved out distinctive work involving the physics of biology. Their work emblemizes the need for such centers devoted to the interfaces of physics and biology.

For Libchaber, Detlev W. Bronk Professor, biology provides an interesting means of studying information. "Life is a singular solution. For me it is the singularity of the solution which is interesting." By this Libchaber, an experimental condensed-matter physicist who studies DNA computation, means that biological idiosyncrasies are just as interesting to him as a physicist as to biologists, only for vastly differing reasons.

At the level of DNA, for example, biologists are interested in aggregations of proteins as "malfunctions" that cause disease. In this sense, the malfunctioning of DNA gives it a machine-like quality that Libchaber and other physicists find useful. Libchaber explains that DNA is essentially a Turing machine, or a computer that consists of a "tape" coded with information and a reading/writing "head" that reads the information coded on the tape. When the head stops reading, it has a "solution." In molecular biology, the ribosome that reads the messenger RNA is the reading head and the RNA is the tape. There are stop codons at the ends of each gene which tell the ribosome when to stop reading and allow it to have its solution, or gene expression.

Libchaber says, "I benefit from participation in biological seminars, but I am not interested in the disease. I am interested in the malfunction. The malfunctioning of such a molecular machine tells you a lot about the machine… I’m interested, for example, in understanding why proteins don’t aggregate all the time."

Speculating on DNA as a type of computer addresses Libchaber’s goal of creating a new machine, modelled on DNA, but one that consists of different materials and conditions useful for computing. Libchaber runs experiments in his lab that test the controls on DNA as a machine.

"In the cell, the control is biochemical," Libchaber explains. "Outside of the cell the control can be different. One study we have published is the effect of turning on and off the heat for control of the molecular mechanisms." Another aspect of DNA that Libchaber studies in his lab is evolution. "I try to change what is written by molecular evolution. I try to eventually make another ribosome, another machine that will not be DNA but will elaborate further the concept of molecular computing."

While perhaps more abstract than the goals that many biologists have in their studies of DNA, Libchaber’s are no less signficant. They are simply directed toward a different end. As he explains, "Computers now, in order to do what they do so fast, need to scale down in size. Right now they are reaching a limit of about .1 microns. If you go below that limit, by a factor of about 10, you enter molecular biology’s scale." Biology offers a model for what computing may be within the next 40 years, as processing speeds double approximately every year.

Professors Eric Siggia (top) and Albert Libchaber are among the Rockefeller University scientists who are combining physics and biology.

Siggia, head of the Laboratory of Theoretical Condensed Matter Physics, occupies himself with molecular biology, too. However, his is a different focus. Siggia is committed to providing the means of generating qualitatively different questions in molecular biology, based on the quantitative work he does with bacteria genomes. His working premise is that understanding more about the non-coding regions of DNA will provide useful information about species evolution and the workings of cells. This is a physical problem as much as it is a biological one. The regulatory parts of the genome are little understood, and biologists for the most part haven’t focused on them in their push to find biomedical applications for gene mutations related to disease.

Siggia’s work takes up where some biologists’ questions leave off and makes use of fully sequenced genomes. Thus far Siggia uses the bacterium E. coli as a reference point, because so much sequence data exists for it. His challenge is to find other bacteria genomes (such as salmonella, cholera, pneumonia) that are close enough to E. coli’s genome, but still different enough to be informative. Then there are three main processes of comparison that can be done and that reveal similar non-coding regions among species.

One method is to use probabilistic models of all the control regions among selected species, which Siggia and his colleagues have done with bacteria. This helps determine which regions might be functional. A more potent method is to mathematically compare the actual genomes of different species, looking for shared control regions and examine their positions and densities. This requires full sequence data. Finally, another method is to analyze gene expression data such as mRNA expression to determine its causes across species. All three methods in combination would generate a full palette of constructive data. The problem is that it is not yet possible to thoroughly achieve all three methods for any set of species yet. With bacteria, Siggia can achieve the first two methods.

Once you generate some reliable data of similar non-coding regions you can start to ask interesting new questions, says Siggia. For example, how does evolution work at the molecular level? What role might shared non-coding regions in a species play in cell regulation?

Another way that Siggia explains some of the premises of his work is that, if genetics is a language his focus is on grammar. What biologists have been studying so far is vocabulary, a crucial part of the language apart from grammar. The way Siggia studies "grammar" is by looking at whether and how non-coding parts of DNA somehow structure the rules, just as the coding parts of DNA, or genes, structure the activities of the language. Bacteria genomes are good to use because there is a lot of sequence data on them, more so than for the mouse or human. Siggia’s work with bacteria genomes have encouraged him to consider other organisms, for which there is also much sequence data. For example, a future collaboration with Rockefeller Professor Fred Cross might lead to discovery of new control regions in yeast. After all, Siggia explains, "in yeast, for example, there’s something very non-trivial going on that we simply don’t understand."