By Branwyn Wagman
Researchers at McGill University in Canada have made a game of the UCSC Genome Browser.
Called Phylo, the game challenges players to create the best possible alignments between sequences of DNA, RNA, or protein to identify similar regions. Such similarities might have resulted from relationships between the sequences in question, either functional, structural, or evolutionary.
McGill University computer scientist Mathieu Blanchette said, “The idea is very simple: it’s a puzzle game where people try to align multiple DNA sequences and maximize the score obtained.”
Phylo is more than just a game, because players are improving actual sequence alignments as they go. The point of Phylo, according to its web site, is “harnessing the computing power of mankind to solve a common problem: multiple sequence alignments.”
Because humans are so good at visual pattern recognition, using Phylo can allow the player to improve on the alignments computers generate, optimizing the computer’s output by sorting. Players sort sequences by color to try for the best match.
Biologists might use such alignments to trace the sources of genetic diseases.
The game became immediately popular, Blanchette said, attracting a few thousand players within the first two days.
The browser group at UC Santa Cruz made alignments available to the McGill group so that the game consists of sections of human DNA thought to be linked to genetic disorders such as breast cancer.
The alignments used as puzzles are short 21-base-pair regions extracted from the UCSC Genome Browser multiple-alignment tracks. The sections being sorted by Phylo compare the DNA of humans to that of 43 other species.
To make a more manageable game, Phylo uses two to eight species at a time. Once a player finishes a puzzle, the results are sent back to the McGill database, where results are combined with those from the other 36 sequences to produce a 44-way alignment.
The McGill team hopes these new alignments will be more accurate than the original alignment. If they are, the resulting optimized data will someday appear in the UCSC human Genome Browser.