By Ben Schiller

Big data is a potential "goldmine" for biomedical research, but handling terabytes and petabytes of material can be a huge challenge. As Tim Thorne, innovation director at Cancer Research UK, puts it: "The problem is not getting the data. It's the time it takes to analyze it, and interpret it."

To speed things up, the charity is turning to citizen science: asking the public to do basic tasks normally performed by professional researchers. Its new platform, called Cell Slider, gets volunteers to look for irregular tissue cells, and classify the brightness of cell stains (indicating the level of a particular protein).

So far, the response has been positive. After 10 days, people have already assessed about 140,000 images. Thorne says the site can do in a week what 10 professional researchers usually complete in a month. "We're looking for ways of speeding up the process," he says. "We might have teams of highly qualified pathologists who do research in their spare time. A lot of that will be churning through data from drug trials, or basic science, looking at lots and lots of sides."

"Their time is precious, and also highly expensive. If we can find ways to break down that process, make it engaging for the public, and produce a similar level of accuracy, it could be very useful."

Cancer Research UK is the second largest funder of cancer research in the world (after the U.S. government). It is currently using Cell Slider for breast cancer samples: The staining indicates the level of oestrogen receptor, or ER. By classifying the molecules, volunteers are helping to assess receptivity to hormone treatments, such as tamoxifen.

There are some controls, as you might expect. Trained researchers look at 10% of the slides first, allowing Cancer Research to compare accuracy levels between the pros and amateurs (so far, there isn't much difference). And two: At least five volunteers look at every slide. Moreover, the charity can put more weight on results from people with better accuracy rates (volunteers register first, so their performance can be tracked).

Cell Slider grew out of a hackathon Cancer Research organized in London earlier this year, and is a joint project with the Citizen Science Alliance, which has several other nice collaborative projects.

Cancer Research is planning a second site next year, called Gene Runner, looking at cancer genomes. Thorne says the charity wants to re-assess many historical drug trials in case researchers discarded treatments that might now be useful.

"The big breakthrough is going to be treatments based on the genetic basis of the cancer," he explains. "There may be cures hidden in the data that were effective for only 10% of the population but which may be effective with 100% of people with a certain genetic profile."

"That's where this sort of research could be particularly useful, because we've got to churn through petabytes of data over the next few years."

Comments