Building molecular systems where computing and decision making are carried out by the chemical processes themselves will have revolutionary consequences in industrial and health applications. For example, cells engineered to produce biofuels will have embedded control modules continuously optimizing their yield. "Smart drugs" will target cancer based on complex functions of relevant indicators without affecting healthy tissue. To achieve this vision, we'll need both: (I) composable molecular parts, and (II) a principled understanding of how to combine these parts into smart molecular systems. I will talk about my efforts on both fronts: First, I have been developing the technology of "DNA strand displacement cascades" that provides programmable molecular parts for functional networks. To address the second challenge and systematically create complex behavior, I am drawing on formal models of computation to achieve systematic and scalable design, as well as rigorous analysis. A new notion of a "chemical algorithm" is emerging that fuses algorithmic ideas from computer science and the basic laws of chemistry and thermodynamics, to enable the rational design of smart molecular systems.
Wednesday, March 19, 2014
Free and open to the public