Skip to main content
Premium Trial:

Request an Annual Quote

The Crowd Cluster


Who needs a supercomputer or a cluster when you've got gamers? Non-scientists, many of whom may not even know the difference between RNA and DNA, are eagerly making major contributions to protein folding, synthetic molecule design, and phylogenetics research, and having fun while doing so. The academic community has released several online games in an effort to solve complex biological problems. This approach is really a combination of crowdsourcing — specifically, taking advantage of the human ability to perform multidimensional pattern recognition, which computers still can't do — and distributed computing, both of which could help researchers solve difficult problems in an innovative, cost-effective way.

The most well known of these games is Foldit, a graphical interface game that challenges players to manipulate a protein's structure and scores their designs based on how well the protein is folded. By learning how humans determine how a protein folds or morphs into a 3-D structure, researchers ultimately hope to refine protein prediction algorithms.

In June, Foldit's developers at the University of Washington published a paper in Nature describing how top-ranked Foldit players excelled at solving challenging structure-refinement problems and in the process helped researchers develop new analysis strategies and algorithms. Foldit's principal developer Zoran Popovic, an associate professor of computer science and engineering at UW, says that Foldit is remarkable not only for aiding investigators with new solutions, but also because it demonstrates that non-scientists who spend sufficient time playing the game can actually become experts in a specific area.

"How [does someone] publish a paper in Nature, who knew nothing about biochemistry before they actually uncovered the solution not previously found in any other way? That requires a certain persistence of time, and games are the most natural paradigm within which that can happen," Popovic says. "Initially when we did Foldit, we had no idea how to present the problem, and the game, in a way that will be conducive for people to discover these novel spatial structures. But over time, as we observed what people are doing, we were able to adapt the computer to best suit the recognition process of people."


Games galore

In January, a collaborative effort between computer scientists and biologists at Stanford University and Carnegie Mellon University released a game called eteRNA, in which players build RNA strands and predict how they might fold or synthesize. The game provides users with a 2-D model of an RNA strand with which they can design complex structures like lattices and switches. The game's lead developer, Adrien Treuille, assistant professor of computer science and robotics at Carnegie Mellon, says the concept of eteRNA was inspired by the success of Foldit, on which he had worked as a developer.

"We were talking about how to apply Foldit to RNA and we realized that [one] of the most exciting properties of RNA is that, unlike protein, it can be synthesized ... and we realized that synthesis and validation can become part of the game. We really thought: 'This is something new and exciting and we have to do this,'" Treuille says. "We wanted eteRNA to work both from a biological perspective to expand the sorts of phenomena that can be crowdsourced essentially, and also from a computer science — and even sociological — perspective, which was how can we organize people to solve complex problems online." Little more than a week after its release, eteRNA had registered 15,000 players who were generating roughly 350 synthetic RNA designs a week, collectively.

Late last year, researchers at McGill University released Phylo, a pattern-matching game that looks similar to a Sudoku puzzle, in that it's made up of squares that represent amino acids. Phylo takes advantage of a player's ability to line up as many of the same-colored blocks as possible in order to identify regions of similarity, while avoiding gaps in the pattern, which represent mutations. In essence, Phylo is like a multiple sequence alignment done by hand — tens of thousands of hands, to be more precise. Phylo's developers contend that because remedying the complex heuristics that algorithms use to conduct multiple sequence alignments would be too expensive in most research settings, crowdsourcing is an ideal alternative. Every successful alignment is analyzed and stored as part of the University of California, Santa Cruz's Genome Browser database, where it will eventually be re--released back into the global alignment as an -optimization.

Jérôme Waldispühl, assistant professor of computer science at McGill, helped develop Phylo. He says that shortly after its public release, Phylo's user base expanded at an incredible rate. "In two months we had more than 10,000 registered users and many more that play without registration, and we've had roughly 200,000 solutions — which means we have probably had 400,000 try and solve the problem. Needless to say, the impact was much bigger than I expected," he adds. "Right now, we're thinking about improving the game, but we would like to apply this approach to other problems related to systems biology, perhaps some classification of data or try to find some signaling pathways."


Phylo is currently only available in French and English, but the team plans to offer the game in Chinese and German as well. They are also working with Nokia to develop a mobile version of the game — which will be available through the company's mobile application store Ovi — in order to further expand the crowd, and therefore the project's computational power. Eventually, Waldispühl says he and his colleagues would like to develop a library of games where users and researchers can choose to address different biological problems. "It turns out that in biology ... there's a huge diversity of problems that we have a hard time to modeling correctly, so games and crowdsourcing allow us to solve problems using a different approach," Waldispühl says. "Instead of using a rigorous algorithm, we can use a more flexible method and rely on human intuition to find nonconventional answers that couldn't easily be found."

In the next year, bio-gamers should expect to see similar games come out of development. Popovic and his team are hard at work on a drug-design game that challenges players to target a virus with inorganic and organic compounds in order to neutralize it, or to determine the best ways to bind compounds to proteins. Scientists will sign up to try out all the resulting drugs in the wet lab for verification and further study. "These types of games are going to change the way a lot of things are going to be done, and we're really empowering people as a whole with a new way of doing things," he says. "The creation process will now be owned by the public instead of pharmaceutical companies."

Filed under

The Scan

UCLA Team Reports Cost-Effective Liquid Biopsy Approach for Cancer Detection

The researchers report in Nature Communications that their liquid biopsy approach has high specificity in detecting all- and early-stage cancers.

Machine Learning Improves Diagnostic Accuracy of Breast Cancer MRI, Study Shows

Combining machine learning with radiologists' interpretations further increased the diagnostic accuracy of MRIs for breast cancer, a Science Translational Medicine paper finds.

Genome Damage in Neurons Triggers Alzheimer's-Linked Inflammation

Neurons harboring increased DNA double-strand breaks activate microglia to lead to neuroinflammation like that seen in Alzheimer's disease, a new Science Advances study finds.

Long COVID-19 Susceptibility Clues Contained in Blood Plasma Proteome

A longitudinal study in eBioMedicine found weeks-long blood plasma proteome shifts after SARS-CoV-2 infection, along with proteomic signatures that appeared to coincide with long Covid risk.