Big Blue's World Community Grid and a University of Washington team are using PCs globally to develop better strains of rice
While the world food crisis demands both quick fixes and long-range solutions, time is of the essence. With food costs soaring, more than one billion people could find themselves at risk of starvation or severe malnutrition in the coming years. So a project announced May 15 by University of Washington researchers and IBM's (IBM) World Community Grid to speed development of better rice may be just what's needed.
By tapping a cluster of nearly 1 million PCs scattered around the world, the researchers hope to develop more nutritious, robust strains of rice sooner by completing complex genetic calculations in just one or two years. Those calculations might have taken 200 years if left to the school's computers. "We can make things happen much faster. We should be able to get new strains to farmers within five years," says Ram Samudrala, associate professor of computational biology at the University of Washington in Seattle.
No Time Wasted
Samudrala heads a research project called Nutritious Rice for the World. The goal is to take genetic information about rice and use mathematical algorithms and three-dimensional computer modeling to discover exactly how the proteins within the rice interact with each other. Samudrala's team provides the information to organizations that crossbreed different strains of rice, creating hybrids that taste good, resist drought and pests, and are more nutritious. By using detailed computer models, researchers can avoid time-consuming and expensive trial and error.
The World Community Grid, launched by IBM in 2004 as part of its corporate social responsibility program, is essentially a dispersed supercomputer that relies on individuals to donate their spare computing power when they're not using their PCs. In addition to coordinating the effort, IBM supplies software that links the computers—be they Windows, Linux, or Macintosh—and schedules computational work.
People who volunteer to have their computers used this way download the software and leave their machines on when they're not being used. The networking software does the rest. When the PC sits idle, the software requests work from a central server. When the PC completes the computation, it sends back the results, and the software requests another assignment.
Breakthroughs in Rice Genetics
The Grid now has such tremendous computational power that it's ranked as the No. 3 supercomputer in the world. Organizations that want to tap in have to prove themselves worthy. So far, there are seven projects underway, including research into AIDS, cancer, muscular dystrophy, and climate change in Africa. Joseph M. Jasinski, who runs the Healthcare and Life Sciences Institute at IBM Research, says Samudrala's team is tackling an important challenge, as rice genetics are a relatively new area of study. "Until you know what the genes are and what the proteins actually do, you don't know enough to make a difference," he says.
The rice project began shortly after The Beijing Genomics Institute in 2000 achieved the crucial breakthrough of mapping the entire rice genome. Samudrala received a grant from the National Science Foundation to study and "annotate" the results. He and his 30 researchers essentially create maps of how rice cells work. He bought 400 PCs and linked them in a cluster, but quickly ran into limitations in the computing power of those machines.
Samudrala learned of the World Community Grid because one of his research colleagues was working with IBM on another Grid project. At first, Samudrala was skeptical that a widely scattered collection of PCs could do the job for him, but he came around. Stan Litow, IBM's vice-president for corporate citizenship, praises people who volunteer their computers for the Grid, and urges others to get involved. "Some people write checks for social causes. Others contribute time. Here's something that doesn't cost you anything," he says.