Programming biological cells

It appears that biological organisms can be harnessed as substrates for computation. Biological cells possess important characteristics, such as energy efficiency, selfreproduction, and miniature scale, that make them attractive for many applications. Examples include embedded intelligence in materials, sensing, smart medicine, and nanoscale fabrication. Vast numbers of programmed cells executing in parallel will enable cheap computation. This abstract argues that individual cells are programmable, and presents a programming paradigm for colonies of cells. Cellular Gates A fundamental chemical process in the cell is the production of proteins from genes encoded in the DNA. The cell performs important regulatory activities through DNA-binding proteins that repress the production of specific proteins. [2] proposes using this regulatory mechanism to implement digital logic inverters. This idea can be extended to construct complex digital logic, making the cell a self-contained computational unit. Microbial circuit design creates a biological digital circuit using a small set of basic gates and a database of protein kinetic rates. To prevent interference between the gates, a different protein is used for each unique signal. Therefore, the number of proteins needed to implement a circuit is proportional to the complexity of the circuit. The design process requires searching the database and assigning suitable proteins to each gate, where the dynamic characteristics of the gate may differ depending on the proteins chosen. The basic gates must be robust enough to function with a wide variety of kinetically different proteins. Figure 1: Genes and Proteins of a Biological Latch We are currently implementing a digital abstraction in E. coli bacteria. A model of the chemical dynamics of digital logic was developed using characteristics of known proteins encoded by the lambda-phage virus. Several basic gates have been designed, such as a biological RS-latch for storing state. The latch uses two input proteins for and, and cross couples the output proteins and to maintain Authors are advised by Gerry Sussman, Hal Abelson, and Tom Knight