Scaling up the complexity, design space, and accessibility
of artificial molecular machines
Natural molecular machines involve three types of key components: structures that provide compartmentalization and spatial organization, motors that perform mechanical tasks such as transportation, and circuits that process information from the molecular environment and make decisions. To build artificial molecular machines that approach the sophistication of the natural ones and that are fully programmable by humans, it is necessary to establish engineering principles that allow the scaling up of their complexity, design space, and accessibility, such that the building blocks and system-level design tools are in the hands of everyone who is interested in exploring diverse applications.
I will discuss recent progress toward this goal through three example systems: DNA robots that sort molecular cargos, micrometer-scale DNA nanostructures that carry arbitrary patterns, and DNA-based neural networks that recognize sloppy "molecular handwriting". For programming DNA robots that perform nanomechanical tasks, we highlight the importance of algorithm simplicity, building block modularity, and collective behavior, which could lead to the development of general-purpose molecular robotics. For programming the assembly of DNA nanostructures, we highlight the importance of hierarchical and recursive strategies, which enable structures with increasing sizes and arbitrary patterns to be created using a small and constant number of unique DNA strands. For programming DNA circuits that perform pattern recognition, we highlight the importance of efficiency and reconfiguration, which opens up the possibility of molecular machines with learning capabilities during autonomous operations. We also highlight the importance of software tools that automatically translate high-level functions to low-level molecular implementations, which facilitate the transformation of algorithmic and architectural foundations to programmable molecular technologies.