COLIN ANDERSON/GETTY IMAGES
Video games designed to tackle tough scientific problems are leading to breakthroughs in RNA structure, protein folding, genetic sequence alignment, and more.
Eli Fisker has struggled to hold down a job as a librarian, largely as a result of an undiagnosed condition he describes as similar to Asperger’s. In his ample spare time, the 35-year-old from Alborg, Denmark, plays an online video game in which he arranges colored discs into two-dimensional chain-link shapes. It’s addictive, and he plays for hours on end. But EteRNA is not your typical Internet time suck: the discs represent nucleotides, and the patterns they form are blueprints for RNA molecules.
Every 2 weeks, the best designs—voted for by the players themselves—are synthesized in the lab by the Stanford University scientists who helped to create the game, and observations about how the resulting molecules behave are relayed to the players. That feedback informs the development of new playing strategies, which in turn help the scientists to better understand the rules of RNA folding and function. Although there’s no PhD after his name, Fisker, one of the best of the game’s 40,000 registered players, is helping to unravel a fundamental aspect of biochemistry that has long eluded the world’s brightest scientists—and even helping to design novel RNAs that encode proteins to fix carbon or fight disease.
© ALEX SLOBODKIN/ISTOCKPHOTO.COM“Our ultimate goal is to create a platform that will enable anyone to design RNAs for anything, and have them tested via our experimental pipeline,” says Rhiju Das, a biochemist at Stanford University who helped develop EteRNA. “We call it cloud biochemistry.”
Launched in January 2011, EteRNA is one of a small stable of video games that enlist the collective intelligence of players—most with no scientific background—to solve fiendishly difficult scientific problems. It’s early days yet, but the approach has already notched some impressive achievements, including a string of widely discussed discoveries published in high-impact journals. It won’t solve every scientific problem, of course, but some argue that the recent “gamification” of science has enormous potential.
“We’re still in the proving stage, exploring how useful [gaming] can be,” says Andrew Su, who runs a bioinformatics lab at the Scripps Research Institute in La Jolla, California. “But I’m very optimistic.”
The first and arguably most influential research game was Foldit. Created by structural biologists and computer scientists at the University of Washington in Seattle, Foldit challenges players to work out the three-dimensional structures of proteins by folding chains of virtual amino acids into optimal configurations. Results generated by online players have already accrued multiple publications in Nature journals since the game’s launch in May 2008.
The game grew out of Rosetta@Home—a project that farmed out computationally intensive protein-folding simulations to home computers. But rather than simply exploiting the spare processing power of PCs around the world, Foldit also makes use of the brainpower of computer owners by framing the problem as a competitive online game.
Players are presented with a hodgepodge of zigzags, squiggles, and loops representing the amino acids of a protein. Moving the cursor allows users to grab, bend, wiggle, and shake various parts of the molecule, with the aim of folding the messy structure into its optimal shape—the form that has the lowest energy—just as molecules tend to do in real life. The more stable the structure, the higher the score.
Foldit players quickly proved they could outperform the randomized runs of Rosetta, which simulates and tests millions of tweaks to the chain to find the shape with the lowest energy. In a challenge to work out the structures of 10 proteins that the scientists had already solved, the players got closer to the true structure than Rosetta for five of the proteins and matched it on three. Then, in September 2011, Foldit players made a breakthrough: they solved the structure of a retroviral protease of the Mason-Pfizer monkey virus, which causes an AIDS-like disease in monkeys—a problem that had stumped scientists for a decade. The study was published in Nature Structural and Molecular Biology (18:1175-77, 2011), listing the “Foldit Contenders Group” and the “Foldit Void Crushers Group” among its authors.
The achievements of the Foldit player community are really eye-opening. They’ve clearly hit on a model that is fun and scientifically productive.—Andrew Su, Scripps Research Institute
The game’s creators have also devised a way to extract and replicate Foldit players’ best folding strategies—which they were invited to encode and share as so-called “recipes”—in addition to their solutions. One such recipe, known as Blue Fuse, which evolved as it spread like wildfire among the game’s elite players, proved even more efficient than the algorithms that drive Rosetta (PNAS, doi:10.1073/pnas.1115898108, 2011). “It was really stunning,” says David Baker, a computational structural biologist at the University of Washington, who helped to create the game. “What they had developed, independently from us, was something better than the best current algorithm we were working on.”
© FRANCKREPORTER/ISTOCKPHOTO.COMIn addition to deciphering and refining natural protein structures, the Foldit community is also trying its collective hand at designing better proteins. In January 2012, for example, Foldit players redesigned an enzyme in a way that sped up a reaction crucial to the production of a variety of drugs by almost 2,000 percent (Nature Biotechnology, 30:190-92, 2012). Researchers are also tasking Foldit players with designing and refining entirely new proteins, such as one to bind to and inhibit the influenza A virus, Baker says. “We’re trying to design potential protein therapeutics, and we’re enlisting Foldit players every step of the way.” Furthermore, Foldit’s designers hope to extend the drug discovery element of the game with a new toolbox of organic subcomponents that will enable the design of novel small molecules.
Foldit has helped to establish online games as a credible source of discovery in computational biology, says Su. “The achievements of [the Foldit] player community are really eye-opening,” he says. “They’ve clearly hit on a model that is fun and scientifically productive. Now a lot of people are paying attention.”
Games for everyone
Foldit’s success has inspired other researchers to exploit the minds of gamers around the world to improve upon existing scientific methods. In comparative genomics, multiple sequence alignment (MSA) is used to identify functional elements of the genome and possible disease triggers. If a particular sequence is conserved across different species, it is likely to have an important function; and if such a sequence is mutated in people with a particular disease, it may be the cause. “MSA is probably one of the most important tools in bioinformatics today,” says Jérôme Waldispühl, a bioinformatician at McGill University in Montreal, Canada.
But the computer algorithms employed to perform MSA don’t guarantee perfect accuracy, so Waldispühl and colleagues created Phylo—an online game that transforms the MSA problem into a simple puzzle that anyone can play. The aim of the game is to improve the sequence alignments of the promoter regions of hundreds of disease-related genes from 44 vertebrate species. The sequences are presented as several rows of blocks, color-coded to represent the four bases of DNA, and players shift the sequences left or right in order to find the best possible match for up to eight different species at a time.
Within 7 months of its November 2010 release, Phylo had more than 12,000 registered users and 3,000 regular players. And they’ve proven themselves worthy: 70 percent of the roughly 350,000 MSA solutions generated by the Phylo community are more accurate than those generated by the best computer algorithm (PLOS ONE, 7:e31362, 2012). “The results returned by the players were much better than what we hoped for,” says Waldispühl. “The human brain has evolved to be very good at recognizing visual patterns, and we can benefit from that.”
For problems where there is some quantity that defines correct answers, then I think games will be powerful.
—David Baker, University of Washington
The game is now available on mobile devices in several different languages. And the team behind Phylo is currently developing an interface that will allow any geneticist in the world to sign in and upload sequences to be converted into puzzles and played by the community. Waldispühl says the idea is to give researchers access to the huge amount of human processing power provided by gamers. “What we want to do is make the best synergy between computer and human,” he says.
A newcomer to the scene is The Cure, a game developed at Su’s lab at the Scripps Research Institute to help find better predictive biomarkers for breast cancer prognosis. Launched last September, The Cure works like a card game in which players assemble a “hand” of five genes from a board of 25 genes pre-selected for their relevance to the disease. The gene set that wins is the one that produces the best predictive model of breast cancer prognosis, as determined by a cross-validation statistical analysis.
Research gaming has even begun to reach beyond the domain of molecular biology. Whale.fm is helping marine scientists advance their understanding of whale communication by asking players to match recordings of whale calls from different parts of the world. The exercise is designed to categorize different whale “dialects” and reveal more about the full repertoire of cetacean language.
© NICHOLAS MONU/ISTOCKPHOTO.COMThis burgeoning interest is a testament to the power of competitive, multiplayer online games to solve complex scientific problems, but how far will it go? Will gamification have a profound impact on the way scientific research is carried out? The answer, Baker says, is likely to depend on the type of question being asked. “For problems where there is some quantity that defines correct answers, then I think games will be powerful,” he says. “But there are other problems that just can’t be posed in that way, and those will be less surmountable with scientific games.”
At the very least, it’s a chance for dedicated gamers like Eli Fisker, most of whom have no formal scientific background, to experience the thrill of discovery. And by encouraging people to engage with complex research problems in a fun and intuitive way, such games could inspire a new type of citizen scientist, says Das—one who may find novel solutions that the professionals have missed. “Wouldn’t it be great if [anyone] could just log in from anywhere and start experimentally tackling the mysteries of life?”
Gameplay can engage unfocused students and teach complex scientific concepts, but uneven access to technology and a focus on standardized-test prep are limiting the use of video games in the classroom.
COURTESY OF WAKE FOREST UNIVERSITYIn January 2012, the Education Arcade at the Massachusetts Institute of Technology (MIT) received a $3 million grant from the Bill and Melinda Gates Foundation to develop an interactive online game to help teach lessons in math and science. Like other multiplayer games such as World of Warcraft, MIT’s The Radix Endeavor will invite players to wander an Earth-like renaissance world, where knowledge is hoarded by an evil leader, and problems can only be solved by reconstructing useful and important ideas in math and science. In an ecology-related task, for example, players go to various locations, such as a “night forest” or an open meadow, and work together to capture and tag birds, learning how to estimate population size and study ecosystem food webs by investigating the animals’ diets. And in fields and farms, players are on their own to crossbreed different varieties of plants and learn about Mendelian genetics along the way.
Although one of the more ambitious educational games so far conceived, The Radix Endeavor is by no means the only one. Interest in games for science education has been gaining steam in recent years, with commercial game makers, game-research labs, and even textbook companies developing thousands of games aimed at teaching basic through advanced concepts in biology, math, chemistry, and more. Many game makers, such as the educational company BrainPOP, offer teachers a compendium of games and simulations geared toward specific grade levels. According to BrainPOP’s website, nearly 20 percent of schools in the United States use their games and services. This shift towards using games for elementary through high-school education is a recent change, says Eric Klopfer, the director of MIT’s Education Arcade and the lead scientist on The Radix Endeavor project. “There was a time where we couldn’t use the word ‘game’ in a school setting,” he says. But a lot of that stigma has gone away, he adds, as research has shown clear social and cognitive advantages for certain types of game play. Indeed, the movement was recently validated by President Barack Obama, when his Office of Science and Technology Policy appointed gaming researcher Constance Steinkuehler as a senior policy analyst to rally political support to speed the implementation of “games for impact” in the classroom.
There was a time where we couldn’t use the word “game” in a school setting.—Eric Klopfer,
Massachusetts Institute of Technology
Many students lose interest in science and math because of the curricular focus on rote memorization of facts in preparation for standardized tests. But games help students engage in scientific concepts through immersive experiential learning, which often commands greater focus and provides greater rewards for success. For example, Steinkuehler’s research showed that students identified as troubled or struggling readers demonstrated dramatically improved reading skills while playing games, able to comprehend higher reading levels. “The texts [of the game] are in the context of solving problems,” Steinkuehler says, and the students were motivated to figure out the meaning. Furthermore, she adds, many games require repetitive practice, which is a key and often undervalued aspect of learning.
Zoran Popovic´, director of the University of Washington’s Center for Game Science, also sees the potential for games to change the way students learn. After he created Foldit, a game that uses the reasoning skills of online players to solve publishable protein structures, he began developing games to help improve math and science education. “Education is not a data-driven science,” but maybe it should be, says Popovic´. Games such as Treefrog Treasure and ReFraction, which Popovic´ designed to improve students’ understanding of fractions and other math concepts, adapt to each player’s level, providing problems that are neither too hard nor too easy. The games also allow teachers to readily track how each student is doing on a particular problem, identify the concepts he or she is struggling with, and see which topics are stumping the entire class. “The game can improve the entire learning ecosystem,” says Popovic´.
There are still major obstacles preventing games from being widely applied in science education, however, says Carrie Heeter, a game designer and researcher at Michigan State University. Specifically, many schools do not have access to the necessary technology to power the games, such as computers and Internet access, and state-mandated educational evaluations based on standardized exams still leave teachers pushing students to cram for the test, rather than giving them time to explore scientific concepts through games and other activities.
“The quality of the [science] test is so abysmal,” in terms of accurately assessing science aptitude, says former teacher Douglas Clark, now a learning researcher at Vanderbilt University. Although the next generation of national science teaching standards, expected to be released by the National Research Council in late 2012, will focus more on depth of understanding and less on a cursory overview of a broad range of subjects, there is no equivalent plan to improve the state tests, Clark says. “Until games are seen by teachers as supporting the goals to which they are held accountable,” it’s unlikely they will be widely employed in the classroom, he says.
The game can improve the entire learning ecosystem.—Zoran Popovic, University of Washington
But Heeter, who was part of a working group convened by Steinkuehler at the White House this July to strategize about how to get high-quality games into the classroom, thinks gaming can become a regular part of science teaching even before the new standards come into effect. Games can be tailored towards concepts targeted by standardized tests, she said. For example, she designed her own game, called Life Preservers, to teach students about evolution and adaptation by trying to save planet Earth from invasion. And teachers can work to better integrate game play into the classroom schedule, splitting longer games into smaller segments or utilizing shorter, puzzle-like games. Classrooms that don’t provide computers for each student can play as a class, with the teacher making the decided-upon moves on an overhead projector. In the future, all classrooms will incorporate at least some game play into their lesson plans, Heeter says. “It’s coming.”
© MARIO ANZUONI/REUTERS/CORBISAtop a platform the size of a bathroom scale, players shift their weight to the left, to the right. In unison, an animated balloon sways on the screen, dodging pointy skyscrapers and collecting floating golden stars. But the Wii Fit-based game, relying on the Wii Balance Board, isn’t just for fun; it also helps stroke victims regain their balance—a difficult task even for standard physical therapy. “If you told me to shift my weight like that without playing a game, I would be really scared, and I probably wouldn’t do it,” an anonymous patient participant wrote in a questionnaire (Top Stroke Rehabil, 17:345-52, 2010). “But when it was in a game, I didn’t really think about how scary it was. I had a goal, and I just went for it.”
The makers of the game, Belinda Lange, Albert “Skip” Rizzo, and their medical virtual reality group at the University of Southern California, are part of a burgeoning field of researchers harnessing the power of technology to create engaging tools that tackle the challenges associated with a variety of ailments, from the anxiety associated with cancer to socialization difficulties that plague children with autism. “This game play, this fun stuff, is not just fun,” says Steve Cole, a professor of medicine at the University of California, Los Angeles. “It can really make a difference in health.”
This game play, this fun stuff, is not just fun. It can really make a difference in health.—Steve Cole, University of California, Los Angeles
Computer scientist Gary Bishop of the University of North Carolina at Chapel Hill got into the gaming field in 2000, after a serendipitous encounter with graduate student Jason Morris, who was developing new ways for the blind to access maps of the ancient world. Morris, whose seeing-eye dog had led him astray on the way to his office, stopped Bishop for directions. Upon hearing the myriad challenges blind people confront, Bishop decided to team up with Morris to build universally accessible maps—using audio and other sensory cues, like the sounds of running water near rivers.
The collaboration quickly snowballed, as Bishop began to learn how challenging even simple activities can be for the visually impaired. His work with Morris inspired him to help design hundreds of blind-accessible games, including a version of the popular Dance Dance Revolution (DDR), which gets kids moving. “Blind kids have higher levels of obesity because they have fewer opportunities to exercise,” says computer scientist Eelke Folmer of the University of Nevada, who has designed similar games, such as a version of the popular game of Wii tennis in which the controller vibrates more intensely when it’s time to swing.
© BLOOMBERG/GETTY IMAGESAnd the benefits of such physical-activity video games are not limited to the blind. They can also be particularly helpful for anyone struggling with obesity, as well as children with autism spectrum disorders (ASD), says psychologist Cay Anderson-Hanley of Union College in New York State. In pilot studies, ASD patients who played exercise games like DDR performed better on cognitive tests and showed less disruptive behavior. “The games are very aerobic and cognitively engaging,” Anderson-Hanley says. “[The activity] can increase blood perfusion to the brain, which could bring about a whole cascade of positive effects.”
Other ASD researchers are hoping to use video games to tackle the social challenges associated with the disorder. Pediatrician and autism researcher Zachary Warren, of the Vanderbilt Kennedy Center in Nashville, Tennessee, for example, is developing games with animated characters that allow children with ASD to practice social skills in consequence-free environments. “We’re trying to create meaningful social worlds where we can work on skills that are really challenging for kids with autism,” Warren says. “One of the tasks is to learn about the most embarrassing things that happened to a friend at school. It might be odd to walk up to someone you don’t know at all and ask,” so the children have to learn how to talk to certain characters, read signals, and get information. And if they commit a faux pas, they just lose a few points, rather than being shunned by an entire peer group. The games can even provide data to help researchers better understand the disorder, Warren added. By monitoring patients during game play, examining what held their attention and how they reacted, Warren’s team is gleaning tips on how to refine treatments as well as the game itself. (For more on games targeted towards ASD patients—and a word of warning against their untested implementation—see “Gaming with Autism” on page 24.)
© JUANMONINO/ISTOCKPHOTO.COMIn other cases, video games may simply serve as an effective tool to help patients learn and cope with their diseases. In the P.E. (Patient Empowerment) game, for example, developed by pediatric oncologist Carol Bruggers of the University of Utah School of Medicine and colleagues, young cancer patients feel like they have control over their health by fighting an evil crab (cancer), using varied arm exercises to manipulate a motion-sensing Sony PlayStation controller. Similarly, Re-Mission, a shooter-style game designed by UCLA’s Cole and his colleagues at the nonprofit HopeLab, allows players to hunt down cancer cells while learning about their illness and treatments, including how to make the most of their treatment and manage side effects. The sense of empowerment that patients get from better understanding their disease and how to fight it can not only improve outlook and mood, it can also help keep their therapies on track: cancer patients who played Re-Mission more reliably took their medicine.
“The great thing about games,” says Cole, “is that if they’re built the right way, they can really make patients feel powerful and in control in ways that they generally don’t when they’re on this conveyer belt of medical care.” —Beth Marie Mole