Science assembling a virtual look at life


Monday, September 11th 2000, 12:00 am
By: News On 6


By Sue Goetinck Ambrose / The Dallas Morning News

Biologists usually take the Humpty Dumpty approach to life, breaking down living cells into tiny pieces to study them one by one.
Now scientists want to do what all the king's horses and all the king's men couldn't – put the pieces back together again.

The pieces, in this case, are thousands of genes and proteins. For years, scientists have been studying the pieces, one or two at a time, to try to get a picture of how life works. But now biologists realize they can't get a true picture of life unless they study all the pieces at once.

It's not easy to study thousands of genes and proteins at once, even for the smartest of scientists. Fortunately, biologists have some help. Computers can follow life's pieces better than any person – let alone all the king's men – can.

"The human mind cannot keep track of more than four or five variables at a time," said Bernhard Palsson of the University of California, San Diego. "You've got to use a computer to keep track of these things."

Computers will turn up solutions to life's mysteries never dreamed of, scientists say. Just as no one knows where the story of Humpty Dumpty would go if someone managed to put him back together, scientists can only guess where the new approach to biology will lead.

"Biology now is where physics was 100 years ago," Dr. Palsson said.

Only a century ago, physicists were discovering laws of nature that led to today's cellular telephones and computers. Biology may be at its own new dawn.

"Who knows what will come out of this?" Dr. Palsson said.

Already, simulating life on a computer has taught scientists a few of biology's secrets. Researchers are beginning to understand why organisms like bacteria and developing embryos are such sturdy machines, able to do their jobs even in the face of adversity. Scientists have new ideas of how genes and proteins in the brain work together to create memories. And by engineering new combinations of genes and proteins, researchers can learn how evolution has led to the life forms on Earth today.

Once scientists get good at simulating life on the computer, all sorts of possibilities will open up, scientists say. Researchers envision using computers to simulate how cells' inner machinery fails in diseases like cancer or diabetes. Armed with such a computer program, scientists can test therapies on the computer before wasting time on a lab rat. Scientists may also discover weak points in the network of genes and proteins powering bacteria that cause human illness, Dr. Palsson said.

Most scientists keep track of gene and protein networks with a pencil and paper. But even simple organisms have thousands of genes, and writing down all the interconnections would take a tree's worth of paper and pencils. So biologists are learning what engineers have known for years – using mathematical equations and computers is a lot more efficient.

Instead of depicting each gene-protein connection with a dotted line, "you write an equation for each step," said John Byrne, a neuroscientist at the University of Texas-Houston Medical School.

"For every gene, there would be an equation to describe that gene and how it relates to other genes," he said. "The computer can solve all the equations at once."

Simulating complex systems with computers is standard practice in many areas, including, for instance, the aerospace industry. Engineers use mathematical equations to describe an airplane's wings, engines and other parts on the computer. With the entire operation of the plane described with equations, engineers can test design changes on the computer instead of constructing a new plane. Engineers can also test how well the plane will perform under different weather conditions.

Biologists are starting to do the same for complicated life forms. Just as the wings and engines help the plane stay on course during flight, genes and proteins help the cell navigate life's challenges. Genes, contained in the double-helix molecule DNA, spring into action when a cell is bombarded by hormones or fighting an infection. Genes arm the cell for its daily trials by helping the cell manufacture protein molecules. Proteins have a variety of jobs; some are even expert at activating or stifling other genes, as the cell requires.

Since genes are responsible for protein production, and proteins can switch genes on or off, even a dozen or so interconnected genes and proteins make for a complicated biological machine. And those machines are starting to look more and more intricate to scientists.

"There's an explosion of information about all kinds of biological processes," Dr. Byrne said.

Researchers have deciphered the complete genetic makeup of dozens of organisms, generating "parts lists" of hundreds or thousands of genes and their corresponding proteins. As more "parts" are turned up, computers can help scientists keep track.

Dr. Byrne used computers to try to explain why the brain remembers better some times than others. When presented with a new mental task, the brain learns best if the task is presented many times, but not too close together. In other words, if you want to memorize something, practice once an hour for 10 hours, not 10 times in one hour.

Based on the genes and proteins known to be involved in memory, Dr. Byrne and his colleagues tried to figure out why the spaced practicing is best. Using mathematical equations to represent the interconnections of genes and proteins, the scientists found what others had previously suspected: If the practicing comes too close together, two proteins – one that turns off a crucial gene and one that turns it on – would cancel each other out. But given enough time between practice sessions, Dr. Byrne confirmed, the protein that turns on the gene has time to take effect.

Armed with a good mathematical model, scientists might be able to figure out what the optimum time is between practice sessions. Or, Dr. Byrne said, researchers interested in medications to enhance good or block bad memories might determine what gene or protein in the memory network could be tweaked by a drug.

The same principle should work on any network operating inside any cells. For instance, a new project headed by Nobel laureate Alfred Gilman, of the University of Texas Southwestern Medical Center at Dallas, aims to simulate networks inside heart muscle cells and disease-fighting cells called B cells.

Ultimately, scientists on the new project hope to come up with a computer simulation of the cells – silicon versions for testing virtual drugs before trying more expensive laboratory experiments.

Computer simulations of bacterial cells have already pointed scientists to targets for new drugs. In a recent paper published in the Proceedings of the National Academy of Sciences, Dr. Palsson reported on a computer model of all the proteins known to keep the common E. coli bacteria functioning.

The E. coli protein network was vast, with a total of 720 links. Dr. Palsson knew ahead of time that some links would be essential and others wouldn't.

"We wanted to know which are the essential links in that network," he said. Dr. Palsson has also found essential links in bacteria that cause gastric ulcers and respiratory infections.

Such links, he said, "are targets for drug development."

But the links that aren't essential, although not good targets for bacteria-killing drugs, are interesting too, Dr. Palsson said. The nonessential links show that the E. coli bacteria are hardy microbes. If one of the nonessential links is damaged, the bacteria can find another way to get by.

This robust quality of life also applies to a more complex organism, the fruit fly, scientists have found.

In a recent issue of the journal Nature, scientists from the University of Washington in Seattle described a computer simulation of a network of genes and proteins crucial for the body to form normally early in a fruit fly's life. After coming up with a simulation, the researchers realized that the network was especially well-designed by evolution. Increasing or decreasing the amount of one protein 10 or even 100 times didn't cause the network to crash.

"It's as if it were a radio that didn't care if it had a 10-ohm or 1,000-ohm resistor in a particular spot or didn't care if you put a transistor in backwards," said George von Dassow, one of the scientists on the study. "This is the kind of property that human engineers aspire to, but they don't achieve it nearly as well as this biological network."

In other words, Dr. von Dassow said, the developing fruit fly may be a biological Timex – able to take a licking and keep on ticking.

Understanding how nature builds biological machines with such resilience could inspire engineers, Dr. Palsson said.

"You might learn how to build the properties into other systems like traffic patterns and telephone networks," he said.

Another way to understand biological machines is to create them from scratch. Scientists working at Princeton University recently reported on a biological machine designed to make bacteria blink like Christmas lights. The researchers introduced three genes into the bacteria. Each made three proteins. One of the proteins glowed under fluorescent light, and the other two proteins controlled the gene for the fluorescent one.

The scientists were able to make the bacteria blink, but in doing so found they do not understand how even the simplest gene and protein networks operate, said Michael Elowitz, now at Rockefeller University in New York City.

Rather than blink with a regular rhythm as expected, the bacteria would oscillate from glowing to not glowing erratically, he said. Other networks found in nature are much more precise.

"It points to the fact that any old oscillator that you put together will not work as well as the real ones," Dr. Elowitz said.

The process of perfecting the "man-made" networks could help scientists understand what went on during evolution, as nature did its own tinkering.

"Hopefully, if you make many of these networks, you can learn about what types of designs are most reliable, Dr. Elowitz said. "You'll confront the same problems evolution has confronted."What are they?

Scientists are starting to simulate cells with computers, just as an aerospace engineer might imitate a plane with a flight simulator.

Who needs them?

Biologists do. Research in biology is generating so much information that scientists can no longer keep track of it all in their heads. But computers can make better sense of all the new data.

What good are they?

With cells, or at least portions of them, represented in computer programs, researchers may eventually test potential medications on hard drives before doing more expensive tests in the lab.