Computer Visionary

Visualization has always been important to science. From Leonardo da Vinci’s detailed drawings of the human body to ball-and-stick models of DNA’s double helix, we’ve always needed to see to understand.

And science today is in the midst of a visual revolution.

Medical researchers now look at 3-D digital models of the brain, and biologists rotate complex molecules in virtual space. Animations help scientists find new drugs and discover the basic processes governing the workings of genes.

And behind the curtains, computer wizards like Dr. Chandra Bajaj are building the computer programs that make it happen.

“Computational visualization is the new frontier, or at least one of the exciting frontiers,” says Bajaj.

Bajaj, a renaissance man himself, straddles the worlds of computer science, biology and physics. In his office, a wide-screened computer monitor displays an animation of a strand of RNA being transcribed by a ribosome within a gurgling, aqueous cell. Models of human bones and organs line the shelves above his desk.

“The visual model has two purposes,” explains Bajaj, professor of computer sciences, “to better our understanding of any 3-D structure and of its role or function.”

He says experimentalists collect data about genes, molecules, cells, human organs, planets and galaxies through traditional techniques like x-ray microscopy, magnetic resonance scans and telescopes. Then he and his colleagues at the Computational Visualization Center build computer programs to create images and animations from these data.

Bajaj, who is director of the center, describes the process poetically as “discerning shapes from images or photographs from shadows.” And though beautiful, the interactive images and 3-D models he produces are not just glitz and glamour. They aid in the scientific process itself, just as models have helped improve engineering challenges, like airplane aerodynamics, for decades.

“We do ‘interrogative’ visualization,” he says. “We’re not just making pretty pictures. We’re using analytical and computational tools to find the right solution. Then, we use visualization to calibrate that solution and sometimes to actually make a scientific discovery.”

One of Bajaj’s research thrusts is using the techniques of computational visualization to identify disease vectors and speed up drug design.

“We’re after knocking-off some of the major diseases if at all possible,” he says. “Of course, what we do is develop the infrastructure tools to do this computationally.”

Bajaj is involved in two major computational tasks: the first is to reveal the target, be it a virus or a tumor cell. Bajaj and his colleagues get data on viruses or cells from electron microscopy techniques and build 3-D models from those images.

“We are in seek mode, and we are seeking as much detail about the target as possible,” he says.

In the second major phase—the destroy phase—Bajaj strives for a therapeutic cure for the target disease. In this case, he has developed computational search techniques to test millions of conformations of drugs with virtual target molecules. The key, says Bajaj, is their ability to test the destroyer molecule virtually, to be sure that it doesn’t bind with, interrupt or affect other important molecules at work in the body. (Unintended side effects from drugs are one of the largest constraints in therapeutic development.)

“Multiple repeated trials are made easier by computers,” says Bajaj. “With all of the data, computer visualization is key. Without that we’d be staring at numbers and just trying to make sense of it all.”Com