CPS: Frontier: Collaborative Research: bioCPS for Engineering Living Cells

Abstract:

Our overall aim in this project is to synthesize desired behaviors in populations of bacterial and mammalian cells. To this goal, we define the basis of a next-generation CPS called biological CPS (bioCPS). The enabling technologies are synthetic biology and micron-scale mobile robotics. Synthetic gene circuits for decision making and local communication among the cells are automatically synthesized using a Bio- Design Automation (BDA) workflow. A Robot Assistant for Communication, Sensing, and Control in Cellular Networks (RA), which is designed and built as part of this project, is used to generate desired patterns in networks of engineered cells. In RA, the engineered cells interact with a set of microrobots that implement control, sensing, and long-range communication strategies needed to achieve the desired global behavior. The microrobots include both living and non-living matter, i.e., engineered bacteria attached to metal plates. We apply the overall proposed computational and experimental approach to test the formation of Turing patterns (i.e., dots and patches of different sizes) in E.coli and Chinese Hamster Ovary (CHO) cells and the formation of liver-bud tissue in human-induced pluripotent stem (hiPS) cells. This Frontier CPS project combines control, formal methods, synthetic biology, robotics, and design automation. It develops novel formal methods and control theoretic approaches to pattern synthesis and BDA. It defines new building blocks for bioCPS in the form of robotic technology and synthetic circuits for control, communication, and sensing. While doing so, we develop theory and computational tools whose applicability go far beyond the scope of this project. We propose a novel, formal methods approach to synthesis of emergent, global behaviors in large collections of locally interacting agents. In particular, we propose a new logic whose formulas can be efficiently learned from quad-tree representations of partitioned images. The quantitative semantics of the logic maps the synthesis of local control and communication protocols to an optimization problem. We contribute to the nascent area of temporal logic inference by developing a machine learning method to learn temporal logic classifiers from large amounts of data. While motivated by verification of gene circuits in the BDA workflow, we define novel abstraction and verification techniques for stochastic dynamical systems. By integrating automatic model construction and verification in the BDA workflow, this project significantly advances automation and standardization in synthetic biology. As a result, it impacts a large spectrum of areas ranging from crop engineering to gene therapy. The research on global behavior leads to a deeper understanding of morphogenesis and developmental biology, and can have a dramatic impact in areas such as amorphous computing, nano-fabrication, and tissue engineering, where such patterns could be utilized to differentiate stem cells into tissues and organs. The work performed in this project is also just in time to provide support for very recent technologies such as tissue and organoid on-a-chip, which hold the promise to produce a paradigm shift in tissue engineering and drug development. It will also impact robotics and manufacturing at the micro scale. The project has a rich education and outreach plan, which includes nationwide activities for CPS education of high school students, lab tours and competitions for high school and undergraduate students, workshops, seminars, and courses for graduate students, as well as specific initiatives for under-represented groups.