Phoenix: A Systems Engineering Approach to Spatial and Temporal Pattern Synthesis in Genetic Systems
Synthetic Biology Engineering Evolution Design SEED
Synthetic biologists create new genetic circuits with informal, iterative design-build-test cycles. Traditionally, this process is performed manually, sometimes with the aid of some software tools, but there have been recent efforts to automate large parts of the workflow to efficiently construct more complex systems. Despite these efforts, existing software tools for different tasks do not often interface well leaving users to manually transfer information between tools sometimes resulting in the loss of information and the introduction of error. Additionally, designers are often concerned with how their systems behave spatially and temporally; however, most existing tools are only concerned with the steady-state function of the system and fail to capture these performance specifications. To address these problems, we have developed Phoenix, a bio-design automation tool that utilizes existing tools for defined sub-problems and incorporates new tools for outstanding sub-problems to create a closed-loop, algorithm-driven design workflow. Phoenix takes a formal performance-bound functional specification describing both the spatial and temporal evolution of a system as input. This specification is converted into a hierarchical design tree by decomposing the specification into layers of abstraction with grammars based on temporal logic functions and enzymatic kinetics. This design tree is then traversed to determine which experiments need to be performed to characterize the design elements in isolation, including considerations for combinatorial design. Once the characterization experiments are complete and the results have been input back into Phoenix, a set of the best predicted designs for the input specification is produced for synthesis. Phoenix has been experimentally validated on some classic synthetic biology model systems, which are used in Phoenix as functional building blocks for larger systems. We are also currently developing extensions to Phoenix for pattern synthesis problems that could have impacts in the fields of cell-fate engineering and organoid engineering in the future.