Online RNA design game garners unexpected interest from nonscientists.
It was a typical lab meeting. The latest RNA synthesis results were in, and they were puzzling. The data indicated deviations from the target fold in unexpected ways despite the fact that all eight designs were highly optimized variants of a previously successful design. It looked like another round of synthesis would be inevitable, and the floor was open for suggestions.
“This is different than the RNAFold prediction, at least using the Turner 1999 parameters.” “Yes, the frequency of the MFE in the ensemble doesn’t track with what’s going on in the leftmost 1x1 loop.”
“Could there be an error in the SHAPE analysis?”
“No, position 50 is highly accessible in several similar designs, so it’s not a one-time glitch. But that means we have no idea what position 41 is bonding to …”
The thing was, I wasn’t in lab. I wasn’t even talking to scientists. I was playing an online RNA folding game called EteRNA, and my fellow labmates consisted of computer analysts, project managers, stay-at-home moms (and one dad), retirees and home-schooled teenagers from around the world. And none of them had any formal education in biochemistry.
EteRNA is a National Science Foundation-funded joint effort between an RNA lab at Stanford University and a computer science lab at Carnegie Mellon University. It’s very slick and well executed; you can dive right into designing RNAs without knowing anything about molecular biology or bioinformatics. It’s all about brightly colored dots, wiggling bonds and getting a higher score than your peers, at least at first. Indeed, 95 percent of the players probably are just substituting EteRNA for their nightly round of Sudoku or Bejeweled.
The other 5 percent
What sets EteRNA apart from other games is the lab portion. Each week, top players create sequences designed to fold into a target shape. They then vote for eight sequences to be synthesized at Stanford and assayed for secondary structure content using the SHAPE primer extension assay (which measures 2'-OH accessibility). Every week, eight lucky players anxiously await the arrival of the experimentally observed secondary structure of their RNA design. When it’s not at all what they expected, they become intrigued, hooked and obsessed. Suddenly, they find themselves spending hours each night in intense online discussions with fellow designers, crunching spreadsheets, looking for patterns and bouncing around ideas for the next round.
In short order, players who had never heard of Turner’s Rules proved that the underlying energy algorithm acted only on nearest neighbors. They created online charts showing all possible permutations of adjacent base pairs and their free energy contributions as measured by exhaustive in-game enumeration. Shortly thereafter, they discovered that bulges and loops could be stabilized with a dangling purine at the 3' end of a helix and that certain tetraloop sequences got a bonus. Most significantly, they concluded that strategies that worked for the canned puzzles did not work for the actual design challenges.
One might think, given the game’s emphasis on points and rankings, that players would hoard their best tricks to themselves – just as the gentleman-scientists of antiquity were loath to publish their most significant findings. But that’s not how events have unfolded. Each new insight is posted, analyzed and heatedly discussed in public online forums. Credit for ideas always is duly attributed. Massive design spreadsheets annotated with musings from the top players are freely shared. Most impressively, this cadre of elite players devotes countless hours showing the ropes to newbies so that they, too, can get synthesized and ultimately contribute fresh ideas to the fold – even if it means giving up their own coveted synthesis slots!
How many of us card-carrying scientists honestly can claim to adhere to the same ethos?
There were obstacles, however, to creating this self-sustaining ecosystem. The small, overstretched team of graduate students responsible for day-to-day operations of the game soon found themselves inundated by requests for more detailed explanations of the underlying science. What does free energy mean, anyhow, and why did designs with extremely negative free energies fail to synthesize in the lab? What is a suboptimal fold? Why is RNA only shown in 2-D? Players were getting frustrated and were quickly losing interest.
So I decided to stop being a passive observer. I linked review articles on folding algorithms, some key primary literature and other bioinformatics tools they could use to analyze their designs. I started logging in late at night to field random questions from curious players about anything and everything, from “What’s a tetraloop?” to “How can RNAs be used to treat cancer?”
It was like pouring gasoline on a fire.
The resulting flurry of activity in the forums and chat rooms proved something I had suspected all along. Ordinary citizens can read and absorb primary literature; they can formulate hypotheses, test them and analyze data. In other words, ordinary citizens can participate in science! They just need to be introduced to an interesting problem, provided with the right tools, and given access to someone willing to answer their questions. EteRNA already provided two of the three.
Providing an outreach opportunity
To be fair, EteRNA is a research project about crowdsourcing to optimize RNA folding algorithms. It was not envisioned as an outreach project, judging by the lack of educators on staff and the lengths taken to shield players from the details of the underlying science. And while such measures were necessary to appeal to casual players, the need to provide detailed explanations to sustain the interest of the most talented players initially was overlooked. This has been addressed as of late, as players now are being allowed to participate actively in improving the game itself. There even is a planned series of Q&A sessions with actual RNA researchers.
Perhaps the real problem is that we are all guilty of systematically underestimating the public’s appetite to be meaningfully engaged in science. Consider that millions of gamers donate CPU time to distributed computing projects like Folding@home despite the pinch of electricity bills. Some actually purchase separate computers solely to contribute more CPU cycles to research they find fascinating. How many of them would jump at an opportunity to participate actively in that science, to have it be more than just a pretty screensaver?
This experience really has forced me to think hard about how we, as scientists, go about fulfilling our mandates to be involved in public outreach. Should we really pat ourselves on the back when we open our lab doors to straight-A high school students from dual-doctorate households or ambitious premeds looking to buff their resumes? And when we dazzle little kids with explosions and freezing flowers in liquid nitrogen – is that education or just entertainment? Is this really what the NSF has in mind when they ask us to detail our broad impacts to society?
Meanwhile, out in cyberspace, stay-at-home moms and college dropouts have arrived at their own, home-grown scientific method for creating RNA designs that work – despite being invisible to traditional venues of science education. Maybe, someday, one of these gamers will consider a career in science. Whether or not that occurs, the NSF – and the public that funds it – has gotten a fantastic return on this investment.
Alan Chen (firstname.lastname@example.org) is an NIH-NRSA postdoctoral fellow in the department of physics, applied physics, and astronomy at Rensselaer Polytechnic Institute.