EVE Online: gamers analyse COVID data
All sorts of people have been getting involved in the battle against COVID, and now there’s a new posse in town - computer gamers! They’re helping out via a project that gets players of a massively popular online spacefaring game to solve puzzles as they play and identify immune cells for scientists trying to study the body’s response to the coronavirus. Eva Higginbotham reports…
Jesse - EVE is a difficult game to describe. You can be a space pirate that goes around stealing or killing other capsuleers in the game. You could be someone who does industry, where you look at the markets, or you mine your own materials, and then you build stuff for other players to use, other players to go blow up, or even yourself. You can be a trade mogul where...
Eva - That's Jesse, an avid player of a game called EVE Online, where as you just heard, you can live out a whole complex life in space. But now while you're absentmindedly mining for resources, or waiting to go through a wormhole, there's a new activity to get your teeth stuck into: analysing data from studies on coronavirus as a part of Project Discovery. Ryan Brinkman explains.
Ryan - The first and probably the most important thing is: we're trying to help scientists in the fight against COVID. So one of the technologies that's being widely used is flow cytometry. What we use flow cytometry for is to look at the different cells that are present in our blood, the white blood cells that are used to both detect infection and to fight infection. We take a sample of a patient's blood, and we label the cells so that they glow with light when they run past the laser one by one. What we label these cells with is on proteins, on the cell surface that we know define the function of these cells.
Eva - These protein markers act as signposts to scientists, telling them what kind of cells are present in a patient's blood. This is important because when your body mounts an immune response to something, a lot of different types of white blood cells get involved, and understanding which immune cells are prevalent, and in what proportions, tells scientists loads of information about how a patient is responding to an infection like coronavirus or a treatment. Scientists can now label up to 50 different proteins on the surface of these cells at the same time. But despite the power of this technique, the thing is...
Ryan - Flow cytometry data analysis sucks. We have 50-dimensional data that scientists are trying to look at. It's just really hard to traverse through this data, when scientists are limited by a two dimensional computer screen, to try and find the cell populations that are important.
Eva - Essentially, flow cytometry analysis is time consuming and laborious. It involves drawing shapes around clusters of dots, where each dot represents an individual cell from a patient's blood; and trying to decide if this dot is a part of that cluster, or is it in its own cluster, or maybe that one, you get the idea. And the fact is, it's currently just too complex and subjective to get a computer to do it all for us. After all, humans have evolved to be very, very good at seeing patterns.
Ryan - People who didn't catch patterns very well were all eaten by saber-toothed tigers millions of years ago, right?
Eva - So what Ryan and his team have done is broken down the 50-dimensional data into lots of two-dimensional puzzles. And that's where EVE Online comes in. These 2-D puzzles are prettied up and uploaded into the EVE Online world, where players can access them as a part of a mini game, like a side project to the main business of hunting down enemy spaceships. And the more puzzles you solve, the better rewards you get within the game. Obviously, I had to give it a go, under the supervision of Páll Grétar Bjarnason, who works with a company that makes EVE Online…
I've actually set myself up with an account, I've made my character, I've got a spaceship. So what I'm looking at is basically...
I could see a sort of graph with an X and Y axis. And there were all of these different coloured dots on the screen. Where there was a high density of dots in a section of the graph, they were all coloured red. And then in lower density, they were blue or violet, like a heat map. Each of these dots is an individual cell. And by separating out the clusters, you're sorting out the cells into potentially different cell types. The tricky bit though is deciding where one cluster stops and another begins…
So using my mouse. I'm just now drawing sort of, lines around what looks like sort of an epicentre of dots. How am I doing with my lines?
Páll - I think you're doing pretty good so far.
Eva - Okay, great. I'm going to hit submit, see how I did. Ah, passed! Well, it says I'm now a trainee data analyst, so I'm pretty pleased with that result…
So it's all fun and literally games, but this is also a great example of what's called citizen science. Jérôme Waldispühl, another lead on the project.
Jérôme - Citizen science is the process of involving the population into the scientific discovery process. And that's what we're doing for the Project Discovery project.
Eva - And it's not just about getting more people involved in science. Having this many eyes on the same data means that the results of the analysis are more robust, and also means that the data can be explored in a much deeper way than if scientists were doing it all on their own.
Ryan - We have really, one-on-one collaborations with many of the scientists providing the data. And they're really, honestly, very excited about the results that the players have to have.
Eva - And it's also great for Ryan and Jérôme's teams, because just in the last few months since the project began they have received over 48 million individual pieces of flow cytometry analysis, which is an excellent basis for them to train artificial intelligence algorithms on how to do this analysis automatically. And with over 170,000 players contributing so far, that number is only going to get bigger. Jesse.
Jesse - I encourage everyone, even if you've never played EVE Online, and you want to contribute something to the coronavirus research, to give it a try. It's got some really cool graphics on to show how the flow cytometry works. It's very simple, easy, fun, and therapeutic to get involved with.
Eva - And for Jérôme, it's about more than flow cytometry, more than training algorithms... more even than coronavirus itself.
Jérôme - Because ultimately, what you want to do is to show people that doing science is not that complicated. Actually, it can be fun. Doing it through a game allows you to remove this mental barrier we put for ourselves at contributing to science, and when you start doing it, you realise that you can make a difference, and we can move forward all together toward this science-based society.