Video game technology vs. cancer

Wake Forest scientists use processor power to to simulate cell changes

CorrespondentMay 21, 2012 

The requirements are daunting for a team of scientists at Wake Forest University, in Winston-Salem, who are helping to work on a cure for cancer: the ability to comprehend, interpret and apply complex concepts and data in a new format; the pursuit or completion of advanced degrees; patience and a tireless work ethic, and being a video game master.

OK, that last requirement isn’t mandatory … but it couldn’t hurt.

Samuel Cho and his team use graphic processing units (GPUs) – the same technology that makes video game images look so realistic on the screen – to simulate how human cells work. The goal is to find new targets for tumor-killing drugs by seeing how cells live, divide and die. What they learn could have an effect on all cancers, Cho says.

The project is a marriage of biophysics and computer science, although this union is experimental. Cho, a specialist in both fields, says, “We’re really not interested in making images. However, we’re interested in performing the simulations. So we repurpose the technology for making the video game images to instead do computer simulations.

“What we found is that the simulations we can perform are much, much faster than the way people can normally do it on a regular computer. Video gamers usually purchase relatively high-end video cards that are usually $500 or so for the latest ones. We buy them off Amazon, put them into our regular desktop computer and then write programs to run our simulations on these graphics cards instead of on a regular computer.”

Cho says the graphics cards’ main function is to “render images and get them on the screen. At a very fundamental level, what it’s doing is basically number-crunching because the location of each pixel and intensity are calculated behind the scenes on these graphics processors. That’s how those pixels are determined.”

The main character for these scientists isn’t from the Legend of Zelda or Super Mario series but rather a key RNA molecule that forms part of the human telomerase enzyme. The enzyme “is only found in cancerous cells,” Cho says.

“The way that it works is, in a normal cell there are these things called telomeres at the end of each DNA string. Every time a cell divides, the telomeres are shortened little by little by little until they get to a certain length, and that’s when the cell knows never to divide again and then it pretty much dies.

“In a cancerous cell, the enzyme telomerase keeps on adding to ends of the DNA, so the cell doesn’t know when to die, and it keeps dividing over and over and over again. That’s basically the definition of what cancer is.”

Learning how the RNA molecule assembles itself, or “folds,” and functions in that enzyme provides new information in researching cancer treatments. Cho says scientists “took this very large enzyme and kept on chopping it up into little bits to see whether it would still function. What they found was that only when you cut out this particular RNA section that the function just disappears. So that’s why it’s a critical part of the enzyme.”

Graphic cards and speed

Cho says his project – three years in the making – is just part of a massive cancer research effort that’s aided by rapidly escalating computer science innovation.

“I’m certainly not the only one who’s using these graphics cards to do computational research,” Cho says. “A good example would be that if you take a look at the world’s top 100 supercomputers, the United States used to have the top supercomputer in terms of raw performance. A few years ago, Japan came out with a supercomputer that was much faster than that of the United States and knocked them off the top of the charts. It was mostly because they incorporated these graphics cards into their supercomputers.

“After that, the Japanese computer was knocked off the list by a Chinese supercomputer. The main reason was that they had also used these graphics cards. Now the United States is about to release their own supercomputer. ... It seems like this is the direction that people will be going in for high-performance computing for quite some time.”

It also helps scientists that the sheer number of games consoles has helped GPUs drop in cost and grow in power: “Five hundred dollars is not expensive when you consider that in our simulations, the speed we get is roughly 30 times” that of the average computer, Cho says. “Imagine buying 30 computers versus buying one computer that has one graphics card. In that case, the economics seem to work out better.”

Cho says graphics cards of different levels are available at different places, including from the companies that make the cards. “If you want a very specialized one for a computing cluster, they have servers that you would have to purchase directly from IBM” or some company that specializes in these kinds of products.

Anqi Zou, an undergraduate student who’s working on the project with Cho, says the games’ widespread popularity has the general public curious about using video cards in scientific computing.

“People familiar with video games pretty much all have heard about NVIDIA, the manufacturer of the GPUs we are using, and are very surprised by this notion of using video cards on scientific computing,” he says. “I think this concept performs as a great hook to get the audience interested in this topic.

“This technology can not only be used in understanding the movement of molecular particles and the cause of cancer and developing treatments (what we are doing now) but also be applied in countless fields such as cosmology, combustion science, financial analysis to name a few. Even as a computer science major, I was very impressed when I first knew about this technology.”

Algorithms, analyses

There can be obstacles when combining computers and science – in this case, developing new algorithms and codes for programming on CPUs because of the complexity of calculating due to the myriad interactions involved. This emerging field makes it essential to be an expert in both biophysics and computer science, as is Cho.

In addition, the relative newness of this process makes it even more crucial that its accuracy be proven and reliable. Zou – who analyzes and compares data outputs generated by the simulations, then documents them – says “my role is important because, given that using the GPUs is still very new to many scientists, people are concerned about the accuracy of the results generated from this new technology.”

Still, the use of GPUs and gaming consoles in medical research has been building momentum for several years, with no end in sight. Since 2007, Sony’s PlayStation 3 has helped researchers in the fight against Alzheimer’s, Parkinson’s and other cancers by supporting Stanford University’s Folding@home project: When a PlayStation 3 console isn’t being used for games, its processors can use spare computing cycles and join a network of machines around the world to simulate folding – the mysterious process by which the body’s proteins assemble themselves.

News & Observer is pleased to provide this opportunity to share information, experiences and observations about what's in the news. Some of the comments may be reprinted elsewhere in the site or in the newspaper. We encourage lively, open debate on the issues of the day, and ask that you refrain from profanity, hate speech, personal comments and remarks that are off point. Thank you for taking the time to offer your thoughts.

Commenting FAQs | Terms of Service