An international research team has developed a biocomputer that uses living bacteria to process information. The system is designed to explore problems that can be difficult for conventional computers, including decision trees that require examining many possible combinations.
The research, involving scientists from Canada, Europe, the United Kingdom and Australia, was announced by McGill University on Nov. 3. A paper detailing the work was published in Nature earlier this year. The project has been in development for two decades, according to the report.
The team tested the system using the Subset Sum Problem, a mathematical challenge that becomes exponentially harder as the number of inputs grows. Researchers translated the problem into a branching graph, then built a small physical structure with tunnels matching its branches. Each fork represented a decision, while each route through the tunnels represented a possible solution.
The researchers used E. coli because its movement through fluid is predictable and measurable, and Vibrio natriegens because it reproduces rapidly. As the bacteria multiplied and traveled through the tunnels, they explored multiple routes in parallel. Automated image analysis tracked their movement, helping the team identify paths that met the problem’s constraints. The researchers also adjusted temperatures to prevent overcrowding.
Bacteria process information more slowly than digital systems, but their ability to explore several routes simultaneously offers a different approach to solving certain problems. Reproduction can also create more processing units within the structure. Dan Nicolau, a McGill bioengineering professor and senior author of the study, said conventional sequential computers are not designed for decision trees and can require substantial RAM for such tasks.
Nicolau said the system could potentially run the video game Doom, but its slow bacterial processors would require a very large biocomputer to host enough bacteria. He also argued that biocomputing could offer an approach to computation that does not require RAM.
Sources: CNET.