Phylogenetics: Tetris meets pedigree

4. May 2012

Humans sometimes understand complex patterns better than a mainframe computer. That's what researchers are now utilising in search of diseases – with a computer game.

Ten years ago the Human Genome Project published the first blueprint of the human genome. The blueprint of the human being consists of three million base pairs. For years, scientists have been sifting through this pile of data in order to find genes and sequences that are involved in the genesis of diseases. In this regard direct comparison with the genomes of other species such as mice or monkeys helps. The idea, using this approach, is to find out which regions are enduring in an evolutionary sense. Where DNA sequences have not changed for millions of years, they are likely to be particularly important. Mutations, shifting of base pairs or gaps in these areas could then be dangerous and cause diseases or abnormalities.

Human psyche works using other programming

So far it has been mostly computers, complete with the help of algorithms and programs, that handle the main part of comparison work between species. Yet the methods for comparison are a long way from providing accurate results. Deviations from rules and complicated combinations are very difficult even for mainframe computers to understand.

The human psyche however works using other programming. In an area where computers struggle to perform and even sophisticated algorithms find themselves at their limits, the human is perfectly tuned, using visual intelligence, to recognise complex patterns and relationships. This quality is now also being utilised by scientists. Researchers working with the computer-biologist Jérôme Waldispühl from McGill University in Montreal have developed a game which merely involves a few clicks and is designed to help in solving one of the biggest problems of comparative genetics.
Knowledge of biology is not needed for Phylo: “The players need not know anything about genetics in order to participate”, says study co-author Mathieu Blanchette. “It is purely a game”. In it, the small squares that come in four colours must be positioned on a horizontal line so that they have as many similarities with those of the series underlying it. With each level comes a new line with new squares, which makes combining more and more difficult. The coloured blocks stand for the building blocks of genetic material: the bases adenine, guanine, cytosine and thymine. Each row represents a different species. The players as such quite incidently make matches of genome sequences between different species. The researchers collect the best results of players in a database, then process them further using their computers and special algorithms.

70 Percent more accurate results

The game, reminiscent of a cross between Tetris and Rubik’s Cube, has already thousands of followers. In the journal PLoS One, scientists describe that in just seven months 12 252 people have registered, about 3,000 play regularly and successfully. The “Phylo method” attains up to 70 percent more accurate results than the computer calculations alone. It is not the first time that scientists have relied on swarm intelligence. While Phylo is dedicated to DNA comparisons, the online game Foldit has since 2008 been determining the three-dimensional structure of proteins. In the journal Nature Structural & Molecular Biology, biochemist David Baker from the University of Washington in Seattle and his colleagues last year reported that they, in just three weeks, using Foldit and two teams of players decrypted the structure of an enzyme in the HIV-like Mason-Pfizer monkey virus. Before this, researchers had been fiddling with it for 15 years. The spatialised form of the protein is now helping them to develop new drugs against AIDS.

One of the first projects in which ordinary citizens and amateur scientists were involved was Galaxy Zoo. Internet users evaluated telescope images for the American astrophysicist Kevin Schawinski. They were supposed to assess whether galaxies are spherical or spiral-shaped. Schawinski’s image recognition software failed repeatedly at this task.

2 rating(s) (5 ø)

Copyright © 2019 DocCheck Medical Services GmbH
Follow DocCheck: