An application launched by the Center for Genomic Regulation (CRG) aims to decipher the genome of breast cancer cells with the help of citizens. The researchers hope that the human ability to detect patterns will outperform the results returned by algorithms.
Thousands of volunteers in more than 150 countries are helping a group of researchers from the Center for Genomic Regulation (CRG) in Barcelona to decipher the genome of cells that are used to study breast cancer. They are doing it through the mobile game Genigma, which the institute launched in February. To date, players have generated valid results for half of the genome. This great collective experiment aims to solve the rest in April. In this way, it will be possible to check if the result obtained by human intelligence is better than the one obtained by applying an algorithm.
Breast cancer reference genome
The underlying scientific problem consists of obtaining a map of the DNA of the T47D cell linethat is, cells that are used in laboratories to study the breast cancer. For example, they are used to check the effect of active ingredients. Knowing their DNA precisely allows us to better understand their mode of action: the molecular mechanisms they set in motion.
Scientists have sequenced its genome, but this operation generates a kind of puzzle. Sequencing returns very small fragments of DNA. Then you have to put them one after another. It’s about a oversized puzzlereaching billions of pieces.
To solve it, the researchers use a reference map of the human genome, which represents the DNA of a healthy person. This helps, but is not optimal. “It would be like breaking a map of Barcelona into pieces and recomposing it using a map from 50 years ago as a reference,” he explains. Marc Marti-RenomICREA researcher at the National Center for Genomic Analysis (CNAG-CRG) and scientific soul of Genigma.
The problem is equivalent to “identifying the abnormalities that characterize the DNA of tumor cells,” he explains. Elizabeth Brogliohead of citizen science at the CRG, promoter of the project.
Does brain beat algorithm?
Indeed, there are algorithms that order the pieces of DNA, maximizing the probability that they are contiguous, but they have some limitations: there are several alternative solutions that yield similar probabilities. And there is always the doubt that there is a best solutionwhich escapes them.
The human eye seems to have a certain facility for ordering small pieces of the enormous matrix that represents this problem. The Genigma team was based on this finding to propose the idea of ask volunteers for help. “The idea was to give a piece of the puzzle to a group of people and ask them to put it in order,” says the researcher.
In the app, this problem translates into a row of 15 pieces, the sequence of which players can move around as they please. When a player rearranges them in a way that improves the ordering, they achieve a higher score. When at least 40 players reach the record score, the reorganization of that bit of DNA is taken for granted.
At first, players don’t know which moves improve the score. This is disconcerting, but the challenge is precisely for each one to discover their strategy. “The brain develops a wide variety of tactics. Some analyze rationally, others are guided by the sounds they hear when the score improves & rdquor ;, explains Martí-Renom. The hope is that this heterogeneity of strategies will find better options that elude the algorithm.
“Getting the reference genome of cancer cell lines is a major problem & rdquor ;, says Solip Park, computational researcher at the National Cancer Research Center (CNIO) in Madrid, not involved in Genigma. “It is interesting that this project has taken a more original approach than trying to improve the algorithm. I am curious to see the results & rdquor ;, she states.
The possibility that the participants give a useful contribution to address a major scientific problem puts Genigma “in the category of the most interesting initiatives & rdquor; in citizen science, according to Pietro Micheluccidirector of the Human Computation Institute in New York, also not involved in the project.
Michelucci cites some successful antecedents, such as Stall Catchersin which participants analyze images of brains with Alzheimer’s; fold.it, devoted to the convoluted problem of how proteins fold; either phylowhich looks for parts of DNA shared between different species.
Since 2018, when the project was conceived, Broglio has carried out working sessions with 120 people from scientists to game developers (including the author of this article).
This participatory spirit has continued after 35,000 downloads of the game. The Genigma team has been updating it based on public feedback. Conversations have even been created on Facebook in which users have shared ideas and strategies.
If the experiment works, the group intends to apply the same strategy to other cancers. “If at the end of April we see that it has worked, we will go ahead with four more cell lines”, concludes Martí-Renom.