The style of genetic computing

Abstract

Cells receive a wide variety of cellular and environmental signals, which must be processed combinatorially to generate specific and timely genetic responses. We present here a theoretical study on the combinatorial control and integration of transcription signals, with the finding that cis-regulatory systems with specific protein-DNA interaction and glue-like protein-protein interactions, supplemented by distal activation or repression mechanisms, have the capability to execute a wide range of control functions encoded in the regulatory DNA sequences. Using a quantitative model based on the well-characterized bacterial transcription system, we show explicitly how various regulatory logic functions can be implemented, by selecting the strengths and relative positions of the relevant protein-binding DNA sequences in the cis-regulatory region. The architecture that emerges is naturally modular and highly evolvable. Our findings suggest that the transcription regulatory apparatus is a "programmable" computing machine, belonging formally to the class of Boltzmann machines. We also expose critical shortcomings of the bacterial transcription systems, which limit the genome-wide adoption of schemes for complex transcription control, and discuss how they may be overcome by the eukaryotic transcription systems.