Saturday, January 22, 2011
Treuille On EterRNA - A Game Played By Humans, Scored By NatureAdrien Treuille has always strived to create computer graphics that make the virtual world come alive, whether by simulating the draft of air behind a race car with his Emmy-nominated Draft Track special effect or the folding of proteins in the online game FoldIt. Now the assistant professor of computer science and robotics has developed an online game that literally comes alive.
Called EteRNA, http://eterna.cmu.edu, the game first enlists players to design molecules of RNA (ribonucleic acid), which biologists now suspect plays a much more important role in cells than previously recognized. Then the best of the player designs are selected each week to be synthesized in the biochemistry lab of Rhiju Das, Treuille's collaborator at Stanford University.
The complex, three-dimensional shape of RNA is critical to its function, so these real-life tests of the designs developed with computer tools are critical. Does the RNA self-assemble into the desired shape, or does it fold at all? The synthesis provides the answer, giving players a score, while also making them part of the scientific process.
EteRNA, supported by a grant from the National Science Foundation, began beta testing last fall and was launched publicly in January. Treuille and his colleagues expect the game will solve some of the mysteries surrounding RNA, but they also hope to show that science can be crowdsourced.
You are using online games to involve a wide array of people in scientific research. Why is this important and why are you doing this?
There are many more people in the world than there are scientists. Scientists shouldn't have a monopoly on scientific discovery. They have historically because there are huge barriers to becoming a scientist in respect to the time required and the money required to have a Ph.D. And not everyone can do that. One of the broader trends of the Internet is to break down those kinds of barriers and we're trying to do it in science. Essentially, we are creating mechanisms through which advanced science is accessible to non-experts, so they ultimately can contribute to science as well.
And a game is the primary mechanism you are using here?
The hope is that you'll want to play this even if you don't care that you're contributing to science. For reasons that we don't quite understand, people want to have high scores. It's almost like a caveman instinct. You want to have more stones in your cave than the next person. It doesn't really matter what those stones are. So what we've seen again and again is if you take an activity and attach a score to it, people are extremely passionate about getting high scores and getting whatever recognition they can get. And there is the matter of social capital, which the game tries to give as well, so the better players are respected in the community. Certainly we know that for many of the players the fact that there is a true scientific substrate for their activities also makes the game exciting for them.
Your new game EteRNA focuses on RNA design. Why did you choose that as a topic for a game?
RNA is one of the trinity of basic molecules that make up life, along with DNA and proteins. We're just starting to understand all of what it does, and it's sort of shrouded in mystery, so it's a very exciting place to be biochemically. But RNA has another interesting property from a game-design perspective. Unlike proteins, it can be synthesized and studied very, very readily. At the moment we understood that, we realized we had the basis for a completely new kind of game in which it would be possible to have experiments inside the game loop and that scores could be based on the results of those physical experiments.
How do novices learn how to design RNA?
One of the main design decisions in EteRNA was determining how much biochemistry the community could handle. We think we've reached a good balance. We've taken the relevant biochemical rules and converted them into a game format that really feels like a game - almost like Tetris.
We have tutorial levels that explain the basic rules of the game, followed by a large number of challenge puzzles that teach people through trial by fire how RNA folds, at least on the computer. Once they've solved those challenge puzzles, players are allowed to go into "the lab," where we really don't know the answers.
At this level, we have an experimental road map that starts off with simple shapes that aren't found in nature. We're then going to introduce more and more complicated shapes. If people are going to ace the early shapes easily, then we can rapidly move through biochemical pipeline. If not, then we'll have to spend more time creating them. Ultimately we want to design RNAs that are biologically interesting.
On this game we have computer scientists here at Carnegie Mellon who are collaborating with biochemists at Stanford University. How did that arrangement come about?
Rhiju Das, an assistant professor of biochemistry at Stanford, and I knew each other back when we were both post-docs in the David Baker Lab, which is a biochemistry lab at the University of Washington. I was working on FoldIt back then and he was working on RNA.
We subsequently talked about applying the FoldIt idea to RNA. Years later, one of my computer science Ph.D. students here at Carnegie Mellon, Jeehyung Lee, started working on that. He really quickly produced this jaw-dropping demo.
We were all at a small conference in Washington, and we all sat down one night and talked about how this could be really fun - now what can we do with this? That's where we came up with the idea that RNAs could be synthesized readily, and so synthesis itself - experiments with actual RNA - could be part of a game.
From that point on, we knew we had something really exciting and new, and we knew we had to try it.
You say that EteRNA is played by humans but scored by nature, why is that important?
It captures the fact that it is a game based on experimental data and nothing has been tried like that before. More fundamentally, most games have arbitrary rules set by humans. But what's exciting about science is that fundamentally it's a game played against rules that we don't know. We're trying to figure out how the knights move and how the bishops move. EteRNA gets right to the heart of that. In some sense we've taken away what might be the man-made rules of the game and we allow people to play directly against nature and see how well they can do.
You say players don't have to care about the science to play the game, but when they play the game, are they becoming scientists and do they act like scientists?
EteRNA presents the Internet community with a set of challenges that are fundamentally scientific. Even though the players are not for the most part scientists, they've responded naturally the way scientists do by forming a community, by discussing, by naming things, and even by having some players who do better, who become sort of experts in different aspects of the field and are respected by others. In some sense the organization of the game community exactly mirrors that of the scientific community. It's a beautiful thing.
What have we learned thus far in the limited time that EteRNA has been online? Are you learning things about RNA?
The players have convincingly shown something that we already knew, in that you can design RNAs that work in principle according to our theories, but don't work in practice. They designed RNAs that some theories said should fold someway and they don't. Then they immediately began hypothesizing why certain things work and why some things don't work. And they have successfully designed RNAs now that do fold properly - completely synthetic RNAs that do fold properly. So in some sense we have our first sort of light into this cavern of why RNAs fold properly. That's sort of our opening now to work with the community and try to really explore this space. We ultimately want to come up with a complete and repeatable set of rules to allow us to synthesize RNAs that fold properly in practice.
By: Byron Spice, firstname.lastname@example.org