Phylo is a Flash-based “game” of comparative genomics, which was developed by two faculty members and a team of students at McGill University’s Centre for Bioinformatics to “harness man’s ability to recognize and solve visual problems efficiently.” For those of us who use bioinformatics to align DNA, RNA, or protein sequence, we typically rely on heuristic algorithms. However, the computing power needed to optimize the multiple sequence alignments get prohibitively greater as base pair number increases. This game was designed to use our visual problem-solving abilities to perform multiple sequence alignments on regions of DNA speculatively linked to human genetic disorders as identified through the Human Genome Project. The sequences of interest were initially aligned to homologous regions from other animals using heuristic algorithms though the University of California at Santa Cruz Genome Browser and stored in the Phylo database. As people play the game, their results are stored and then re-introduced to attempt a collective optimization of the global alignment of the entire genome.
As a Flash-based game, Phylo plays very intuitively. There are many levels to this game and at the start, the computer randomly selects a region to align. You may select a specific region by its database ID, or you may select regions based upon the organ systems that a genetic disorder affects. You may also choose to align as few as three or as many as eight sequences. The developers have provided a tutorial to point out the nuances of the game, which is very helpful. Anyone can play the game as a guest; however, if you want to track your alignment history or have access to the “Challenge” section to play against friends, you must register on the site or use a Facebook account.
Phylo is timed and it becomes more challenging the more sequences you attempt to align. The program is designed to have you start with two regions to align at a time. Once you have found the “best” or “par” visual alignment, additional regions are then introduced. You align these new regions to each other and then to the initial pair. The criteria used to determine what is “par” are based on the number of matches minus any mismatches or gaps needed for the alignments. To continue the game, you must at least match the “par” score for the initial heuristic alignment and any optimization provided by other players. If you cannot reach the par score in the time allotted, the game ends. The program also provides the regional sequence information for a related common ancestor at the branch of the regions being analyzed. You can call this common ancestor up on the game page by clicking the phylogenetic tree root to help with your own alignment.
The game presents a very good visual representation of heuristic algorithms and how they process and organize large amounts sequence data. I think it can be used as a good pedagogical tool to visually demonstrate and put into practice sequence alignment without having to get too heavily into algorithm design. The fact that this is all web-based means there is no need for any special bioinformatics software or access to a supercomputer. In my opinion, it is an activity meant for college-level students or extremely bright high school students. Phylo is challenging and students should probably have an introduction to bioinformatics and human genetic disorders before playing. I feel that this background information will help energize the students, especially if you point out how their alignments will help optimize the information put forth by the Human Genome Project. Speaking as someone with a strong competitive nature, my interest was piqued by the more complex alignments. I did fairly well playing different levels, but I could get extremely frustrated with the timed nature of the game. Some younger students could become too focused on the time allowed, rather than the learning benefits of multiple sequence alignment that the game provides. However, players with the maturity will see Phylo for the challenge that bioinformatics can in itself be.