Crowdsourcing cancer research: a smartphone game to analyse cancer gene data

6 Feb 2014

by ecancer reporter Clare Sansom

The leading research charity Cancer Research UK (CRUK) has launched a mobile phone app that allows people with no knowledge of cancer to take part in its research simply by playing a game.

This game, Play to Cure: Genes in Space, is freely available for smartphones and tablets running the Android and Apple operating systems from Google Play and the Apple App Store respectively [1].

Launching the game, Hannah Keartland, citizen science lead at CRUK, explained that the charity’s researchers were generating enormous quantities of data from individual patients but that its analysis was creating a real bottleneck.

People, even people with little or no background knowledge, are still better than computer programs at recognising patterns in this type of data.

“We couldn’t analyse our data automatically, and we can never recruit enough data scientists to do it in-house, so we decided to ask the public to help us with our analysis”, said Keartland.

The first crowdsourcing application developed by CRUK was Cell Slider, a computer game in which users are asked to identify and pick out cancer cells in biopsy slides.

Since its launch in October 2012, over two million cell samples have been classified, with over half of all analyses coming from players based outside the UK.

Data analysis that would have taken eighteen months can now be done in three, and the game is easily adaptable so it can be used with samples from different cancer types.

Following the success of Cell Slider, the citizen science team at CRUK decided to harness the popularity of the smartphone as a gaming platform and develop an app with which game players could analyse genetic data from cancer patients.

The first step was a Game Jam, in which hackers from Google, Facebook and Amazon, academic gaming experts from City University London and some of Cancer Research UK’s own scientists gave up a weekend to develop ideas for turning a data analysis problem into a playable game.

In only 48 hours, the Game Jammers developed 12 prototype games, and these ideas were handed over to an agency, Dundee-based Guerilla Tea, to develop the product.

The data behind Genes in Space is taken from a ground-breaking analysis of genomic data from 2000 patients with breast cancer by Carlos Caldas at the Cancer Research UK Cambridge Institute and his group that was published in Nature in 2012 [2].

This research subdivided the cancers into ten essentially different diseases based on their gene expression profiles, with different prognoses and different optimum treatment regimens.

There is still much that could be learned from a detailed analysis of the genetic defects that characterise each breast cancer subtype, and the over two million data points obtained from each patient in the study form a dataset that would be far too time-consuming for experts to analyse by hand.

“We handed over the copy number data from this dataset to the game developers, stripping out SNP data as this could potentially be used to identify individual patients”, said Caldas.

This data forms a trace along each patient’s chromosomes, and identifying the peaks and troughs in this trace that represent copy number changes will help scientists identify which genes carry the mutations that drive the development of each breast cancer subtype.

The game itself is set in the far future, and revolves around searching space to collect the precious, elusive “Element Alpha” which is needed for medical research.

Players first plot their courses through space and then drive their spacecraft to collect as much Element Alpha as they can; the courses traced during both steps of the game are mapped back to the original genomic data, indicating where players have – unknowingly – identified changes in copy number.

Other game elements, in which players design their spacecraft, dodge asteroids and move up through levels, have no research relevance but are simply introduced to make the game more exciting and playable.

“Playing this engaging game – even if you only have five minutes to spare – will help us identify new patterns in our data, develop more diagnostic tools and personalised medicines, and beat cancer sooner”, said Caldas.


[1] Cancer Research UK, ;

[2] Curtis, C., Shah, S.P., Chin, S.F., 28 others and the METABRIC Group (2012). The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature 486(7403):346-52. doi: 10.1038/nature10983.