Safety-Aware Cyber-Molecular Systems

Molecular programming uses the computational power of DNA and other biomolecules to create nanoscale systems. Many of these envisioned nano-systems are safety-critical, such as diagnostic biosensors that detect contam-inants, drug capsules that dispense medicine when they encounter diseased cells, and configurable nano-robots. Challenges to the safety engineering of the nano-systems include their probabilistic behavior, their very small size, the very large number of them that execute at once, and the dynamic envi-ronment in which they operate. Designs need to assure safe outcomes from highly fault-prone devices.

This poster reports three new contributions by our interdisciplinary project team. The first is the design of a molecular fault detector that monitors the health of another molecular system at runtime and triggers an appropriate recovery action if it fails. The second is the development of robust chemical reaction networks whose behavior and correct output are resilient to small perturbations in their input signal and three other parameters. The third is the design of a molecular device that performs state logging of the modeled system on demand. Details appear in the referenced publications.

The scientific impact of this work is to provide a framework for apply-ing computational and software engineering methods to create safer cyber-molecular systems. The broader impacts of the project include improved support for developers to design safety into these systems. The educational impacts of the project for education include training a diverse group of stu-dents in this new field. We anticipate that the results of this project can extend to other distributed, massively parallel cyber-physical systems that must operate safely in partially understood contexts.

This research was supported in part by National Science Foundation Grants 1247051 and 1545028.