A team led by McGill University has developed a biocomputer - a computing device in which living bacteria perform the role of "processors." According to an article published in the journal Microsystems & Nanoengineering, such a system is capable of solving large, complex problems while using only a small fraction of the memory and resources compared with standard computers built on semiconductor elements.
An ordinary computer, the authors explain, goes through possible solutions sequentially, one after another. For complex problems this can take years or even centuries. Bacteria, by contrast, are able to explore many options simultaneously. In addition, they divide, creating ever more biological "processors."
"Modern electronic computers perform millions of operations per second using a single chip," says Dan Nicolau, senior author of the work and a professor of bioengineering. "Thanks to bacterial division, we have a million chips working at the same time. The more chips there are, the faster a problem can be solved, even if each individual 'processor' works more slowly." According to him, the exponential division of bacteria can scale to the size of the problem: "This is a completely new way of computing."
The project was the result of 20 years of collaboration among researchers from Canada, Europe, the United Kingdom and Australia. To build the biocomputer, the scientists translated a classic problem into a graph of possible choices and then recreated it as a network of tiny connected channels through which bacteria could move. Each route in the network corresponded to one possible solution, and special intersections - "logic gates" - served as decision points.
Rapidly dividing bacteria, moving through the channels, collectively explored many routes at once, testing a large number of possible solutions in parallel. To avoid overcrowding, the researchers regulate temperature, controlling bacterial growth. By observing where the bacteria move and using automated image analysis to track successful routes, the scientists were able to determine combinations that satisfy the conditions of the problem.
The authors emphasize that the biocomputer is not intended to replace classical electronic or quantum computers. Its goal is to offer a resource-saving and cost-effective tool for solving problems that previously seemed unsolvable. "Our approach is designed for problems for which electronic computers are not intended, such as determining protein structure," Nicolau notes. "It has been reported that artificial intelligence has handled this, but that is only half the truth, since it can only review available data. We are offering a technology for solving problems, not for making assessments based on past information."