Cells Are Plausible Targets for High-Level Spatial Languages

Abstract

High level languages greatly increase the power of a programmer at the cost of programs that consume more resources than those written at a lower level of abstraction. This inefficiency is a major concern for the programming of biological systems: although advances in synthetic biology are beginning to allow bacteria to be programmed at an "assembly language'' level, metabolic and chemical constraints currently place tight limits on the computational resources available. We find, however, that the semantics of the Proto spatial computing language appear to be a good match for engineered genetic regulatory networks, and particularly for describing the spatial differentiation necessary to construct tissues or organs. In this paper, we propose a mapping between Proto programs and standardized biological parts. We then demonstrate the plausibility of this mapping by applying it to a band detection program, finding that standard code optimization techniques can transform the inefficient program produced by the initial mapping into an efficient design equivalent to the Weiss laboratory's hand-designed band detector.