Seeing Nature and Data in New Ways

Scientific Visualization Reveals Unexplored Nature, Makes Sense of Complex Data
Share Share This
Jan 13, 2008

Like security camera footage from a gas station robbery, a short, blurry, stop-action movie provided stem cell scientist Tannishtha Reya with a clear view of two well-known cancer-promoting proteins caught in the act.

The camera in this case was a system that captures images of dividing stem cells lighting up fluorescent green in response to a particular gene's activity.

Stringing together a series of these images, Reya's lab created a movie showing that one protein was speeding up cell division. A second protein then forced the dividing stem cells to stay immature, rather than ripening into various kinds of blood cells. Together, they were taking steps that produce a particularly deadly form of leukemia for which treatment options are few.

"Until you could see this, you couldn't know what was really going on," says Reya, an assistant professor of pharmacology and cancer biology in the Duke Medical Center. "You had this output, a very aggressive leukemia, but you didn't really know how it got there. As we watched the movies, we realized that each protein had a unique consequence. Visualization was key." (Watch an example of the research video.)

Such is the power of scientific visualization, a marriage of advancing technologies and evolving mindsets that is revolutionizing scientific discovery across almost all areas of science at Duke.

Some researchers are pushing the boundaries of physics, biology and computing to capture glimpses of nature at increasingly smaller and faster scales. Others are turning to imaginative new forms of visualization that make sense of bewildering equations, huge data sets and complex interactions. (See a slideshow of examples.)

Computers Aid The Vision

"Astrophysics, chemistry and physics have been very computational for many years," says Rachael Brady, who has made a career of helping others visualize their data as director of Duke's Visualization Technology Group (see profile). "Biology and the social sciences are now coming on board with computational models. All of a sudden they can do visualization. They can make these very illustrative graphics that help them convey what is going on.”

Biochemistry graduate student Jeremy Block shares this new experience of biology at the molecular scale with undergraduate students in the DiVE, a six-sided projection of scientific data where computers create the illusion of a 3-D world for people standing inside wearing special glasses.

"They're awestruck," Block says, as his work surrounds him like a glowing galaxy. "It takes them a couple of weeks to calm down enough to begin doing something interesting. They're too busy saying 'this is really cool!'"

Getting inside his 3-D molecules makes things clear, Block says. "You get the sense that you can see a solution quicker and better. You can literally turn your head around a problem, or grab it and pull it closer to you."

The DiVE also has become a classroom for brain scientists Scott Huettel and Michael Platt, who teach neuroanatomy in 3-D.

Hovering before their students, a very large brain can be rotated in space and separated into its various color-coded parts by pushing buttons on a special wireless mouse. Students can also seem to slip inside tight hidden passageways of the brain.

"It is utterly and amazingly clear," exclaimed goggle-clad Huettel, an associate professor of cognitive neuroscience. "I'm always struck by its beauty."

Plant "These structures have a complex three-dimensionality that is virtually impossible to appreciate from the 2-D MRI images we normally work with," added Platt, an associate professor of neurobiology. "I myself have learned anatomy in here to a level I hadn't known before."

Nature Photography, Writ Small

Anatomy at the very smallest scale -- the parts of a cell -- is being seen for the first time with a different kind of instrument. A confocal microscope relies on fluorescent dyes and carefully tuned laser light to take colorful images of cell structure and function in tightly focused slices.

Researcher Uwe Ohler is using the confocal microscope to take what amounts to time-lapse movies of genes switching on and off in a living plant root. Side-by-side “lineups” of different images of the same kind of root "allow us to look at how gene expression changes,” Ohler says. “If I increase the salt concentration, for instance, which genes are now active in a different pattern than they were before?

“Pictures definitely help,” adds Ohler, an assistant professor of biostatististics and bioinformatics at Duke’s Institute for Genome Sciences and Policy (IGSP). “We humans are pretty visual beings.” Ohler's group is now using three-dimensional imaging, taken in collaboration with biology professor Philip Benfey's lab, "so we could look at the roots from different angles,” he says. “And in movies taken of different live roots, we can also stop or fast-forward the roots to see where they are different and where they are the same."

Bucking the 3-D trend somewhat, physics professor Robert Behringer examined how pressure accumulates in packed granular material, a 3-D problem, by using a 2-D model of transparent plastic cylinders that change colors when squeezed. The resulting pictures show irregular "force chains" of pressure snaking through the material.

"It provides immediate, intuitive insight into granular behavior," says Behringer. "We can also take those images and process them to obtain detailed quantitative information. At this time there is no other way to obtain information that is near comparable."

Turning Numbers Into Pictures

In addition to taking data from pictures, Duke scientists are taking pictures from data, turning abstractions into images.

A rapidly expanding field called computational biology relies on elaborate diagrams to make some sense of the fantastic complexity of genes, proteins and signaling chemicals interacting within the cells of plants and animals. These “network models,” depicting how genes are turned on and off and in what order, look a lot like integrated circuits. Map

Much of the data created by environmental science also can be more easily interpreted with visualization. A group led by Marie Lynn Miranda, an associate professor of environmental sciences and policy in the Nicholas School of Environment and Earth Sciences, turns layer upon layer of data about environmental exposure risks and health outcomes in communities across North Carolina and beyond into colorful maps that highlight trouble spots.

Visualization of complexity and abstraction is nothing new to structural biology pioneers David and Jane Richardson, but their tools are getting better all the time.

Working as a team, the Richardsons have given scientists and students alike a much better understanding of the complex 3-D architecture of proteins by inventing something called a ribbon diagram. David charted how X-rays interact with all the atoms in a crystal of protein molecules; Jane translated those complex readouts into patterns of ribbons, lines and arrows that are simultaneously accurate, understandable and beautiful (see profile). Their student, Jeremy Block, is now taking those ideas to the next level in the DiVE.

"A picture is worth more than 1,000 words here," says G. Allan Johnson, director of Duke's Center for In Vivo Microscopy, where data from ten different imaging systems are melded to create breathtaking anatomical images of mice and other small animals. The images show the insides of feathery lungs, beating hearts, developing embryos and “computer-dissected” sections of brains without destroying the animal.

They also create "4-D portraits" of a living animal's cardiovascular system, showing it virtually sliced and changing over time.

"I tell (colleagues) that the amount of data per unit of time that can be acquired now in any study is two, three or four times more than just 10 years ago," Johnson says. "If we gather that much data and have no way to appreciate it, we are wasting a lot of ours and taxpayers’ time."

"Without an understanding of the visualization tools that are now available to us, it's unlikely we'll be able to fully appreciate the depth of the data we have at our disposal," Johnson says.

Produced by: Monte Basgall, senior science writer, Duke News and Communications.


(919) 681-8054