Researchers have achieved faster, more reliable logic operations using DNA molecules to perform computing logic tasks that might one day help diagnose and treat disease. 

Several DNA properties can be repurposed to achieve computational logic. DNA stores information in its molecular ‘base sequence’. A DNA strand can recognize and bind to a matching strand with a complementary sequence. DNA strands can be copied by enzymes, and can have chemical groups attached that can be activated or deactivated by the addition or removal of short complementary strands.

Duke University computer scientist John Reif, and colleagues in the USA and Egypt, exploited such properties to develop fast, compact logic systems equivalent to the OR and AND gates of conventional computing. OR gates send an output signal if one or both of two possible input signals arrive, while an AND gate does so if two selected signals are both present.

In the team’s DNA computing process, the inputs are short DNA molecules that interact with other DNA molecules and enzymes to generate an output signal by activating fluorescence from a chemical group attached to a signalling DNA. The input and output molecules can represent numbers, or more complex aspects of whatever system is being analysed. 

The procedure uses an enzyme that works with single strands of DNA, rather than the double strands used in more cumbersome techniques. A natural enzyme called DNA polymerase is used to copy the DNA molecules up to the levels required.

“We used our simple logic gates to implement a large circuit that can compute the square-root of a four-bit number,” says Reif.

The innovation simplifies and refines the molecular processes to achieve the computation in minutes, rather than the hours required by previous research; and with fewer undesired side reactions.

“The work nicely combines the programmability of DNA interactions and the efficiency of enzymes to provide a robust, simple and fast implementation of molecular logic circuits,” says biomolecular computing expert, Yannick Rondelez, of the French National Centre for Scientific Research, who was not involved in the study.

Andrew Phillips, head of the Biological Computation Group at Microsoft Research in Cambridge, UK, also not involved, describes the work as “an important alternative to more complex DNA architectures.”  

Reif acknowledges that the research still has a long way to go before it can deliver practical applications. But it may eventually allow networks of DNA molecules and enzymes to detect and analyse the molecular indications of disease, and compute and control the steps necessary for the most effective release of drugs.