— A grid of 9 wells corresponds to the squares on a tic-tac-toe grid (Image: Joanne Macdonald)
A computer that uses strands of DNA to perform calculations has mastered the game tic-tac-toe.
MAYA-II, developed by researchers at Columbia University and the University of New Mexico in the US, uses a system of DNA logic gates to calculate its moves.
A DNA logic gate consists of a strand of DNA that binds to another specific input sequence. This binding causes a region of the strand to work as an enzyme, modifying yet another short DNA sequence into an output string.
Scientists have already developed DNA computers capable of various similar simple calculations. But the researchers behind MAYA-II say their design should prove particularly useful for exploring ways to identify the genetic markers associated with certain diseases.
A human plays MAYA-II by adding a DNA sequence that represents their chosen move at a particular point in the game. This is added to all 8 wells that correspond to the outer squares on a tic-tac-toe grid. One limitation of the system is that the human player must always go second, after the centre square has been filled by the machine.
The previous version of MAYA, unveiled in 2003, was even more limited. The human opponents first move was restricted to one of two squares (see First game-playing DNA computer revealed).
Each well contains between 14 and 18 DNA logic gates. After a human player makes their move, MAYA-II responds through a DNA reaction. The strand outputted feeds into a series of other DNA logic gates that link the different wells. This results in a chemical reaction that generates a green fluorescent glow in the square MAYA-II selects as its next move. The strand also interacts with the remaining wells, priming them to respond appropriately to future moves.
"MAYA-II moves bio-computation up to the next level of power," says Joanne Macdonald, a researcher at Columbia University, who helped build the system. "It's similar to the invention of the first microchips with hundreds of logic gates."
"Those kind of experiments are the right direction for DNA computation," says Martyn Amos, an expert at Manchester Metropolitan University in the UK. But he adds that, "it's not something you can pit against silicon".
"Playing with MAYA-II takes a long time," Macdonald admits. The system needs between 2 and 30 minutes to compute each move and a second machine is required to translate the fluorescent signals generated each time into a move in the game.
However, Macdonald believes the technique could help researchers refine techniques for analysis of DNA samples. She is already using DNA logic gates to separate viruses and detect particular combinations of DNA mutations.
Amos agrees that MAYA-II could help researchers in this way. But he warns that the DNA molecules used cannot be controlled perfectly, and could be prone to the occasional malfunction.
Journal reference: Nanoletters (DOI: 10.1021/nl0620684)