Researchers at Maynooth University in Ireland have built a molecular computer that performs 100-bit calculations in a test tube, powered by nothing but heat and a saline solution. According to the university, the work has been published in the journal Nature.

The setup is radically simple. A small drop of water and salt goes into a test tube together with short pieces of DNA and a longer DNA strand that acts as a scaffold. Heating and cooling cycles then drive strand-displacement reactions — the strands bind, unbind and rearrange — and those reactions implement the logic of the computation.

Using this approach, the team ran ten programmes that included 100-bit calculations, working with numbers between roughly 11 million and 34 million. Some runs took up to 14 hours. Crucially, the system proved reusable: it executed up to 25 calculations in succession, a step beyond the one-shot demonstrations that have characterised much molecular-computing research.

Raw speed is not the point. A silicon processor finishes the same arithmetic in nanoseconds; the DNA computer needs hours and a heat source. But it requires no continuous electricity supply, which is what makes the result interesting for settings where power or hardware cannot be supplied — molecular sensing, diagnostics inside a biological environment, or logic that runs where a chip would be impossible.

DNA computing has been promised for decades, ever since Leonard Adleman's 1994 demonstration that strands could solve a small travelling-salesman problem. Most follow-ups stayed at the level of toy logic gates or needed continuous enzymatic or electrical input. What matters here is scale and reusability rather than throughput: the researchers report data sets far larger than earlier molecular logic operations, in a system that can be reset and run again.

The caveats are worth stating plainly. Fourteen hours for one calculation is not a general-purpose replacement for a CPU, and the practical applications — sensing, programmed chemistry, molecular memory — remain speculative. But as a demonstration that biological molecules can carry out substantially more computation than previously shown, the result marks a genuine step up for the field.

Sources