Continuing progress in the miniaturization of electronic devices is driving computer technology inexorably towards molecular scale devices. In just a few more "Moore's Law Cycles," components will approach the size of individual molecules, at which point new computing architectures need to be found. Importantly, the problem is not just to develop molecular-scale computational logic, but to interface the logic to its environment, namely a molecular-scale environment.
Molecular logic circuits may one day be at the heart of embedded chemical controllers for biochemical, nanotechnological, or medical applications --- environments that are inherently incompatible with traditional electronic controllers. Performing computation inside living cells offers life-changing applications, from improved medical diagnostics to better disease therapy to intelligent drugs. Due to its bio-compatibility and ease of engineering, DNA is an ideal physical substrate for carrying out molecular computation. However, current DNA circuits are not fully modular and have not yet been demonstrated to operate in living cells. An approach to DNA computing in which all circuit elements are co-localized on a DNA nanostructure can help address both challenges.
DNA circuits that rely on the strand displacement mechanism constitute the biggest rationally designed molecular circuits by far. However, to make this technology useful for practical applications two major challenges need to be addressed. First, truly composable DNA need to be developed. In current DNA logic circuits all components diffuse freely in solution and encounter each other at random. Whether two components --- the logic gates and the signals connecting them --- react with each other depends on the chemical sequences of the components, rather than their location. Therefore, to compose a circuit with multiple gates, each signal and logic gate must be built with a different set of DNA sequences to avoid interference between gates or modules. The need to specifically design sequences of interacting components limits composability and forms a major hurdle in scaling up the size of DNA circuits. Second, architectures need to be developed that are well-suited for in-cell computing. So far, no complex DNA circuits have been demonstrated to work reliably in cells: the delivery of circuits with many independent components to living cells is challenging, and the relatively simple single and double-stranded components from which existing circuits are assembled are easily degraded by cellular nucleases. A novel approach to DNA computing in which all circuit elements are co-localized on a DNA nanostructure will be used to address these seemingly distinct challenges.
