2013.6.20 Scalable Analog and Digital Platforms for Biological Computation
Timothy K. Lu，Ph.D.
Department of Electrical Engineering and Computer Science and Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA USA
Biological computation is critical to synthetic biology’s goal to implement complex systems that are capable of sensing, processing, and actuating. Digital computation, where signals are abstracted into ‘0s’ or ‘1s’, is applicable in situations where high precision or decision making is desired. The analog paradigm, where continuous variables are represented in continuous physical signals, uses the inherent mathematical equations implemented by natural systems to perform complex computations. In electronics, digital is the dominant computational paradigm since billions of transistors can be assembled together to implement defined, reusable computational devices. Analog circuits are used for specialized applications, such as interfacing external signals to electronic signals, very-low-power computation, and power management. In contrast with electronics, biological systems are highly resource-constrained environments in terms of parts, energy, and wiring. Thus, the digital abstraction may not be the ideal computational paradigm for many biological uses. Analog computation may be more efficient for achieving complex cellular computations in certain applications.
We propose that both digital and analog computation can be achieved in living cells using synthetic biology. Here, we describe our efforts to implement these paradigms with scalable platforms. We have engineered libraries of synthetic transcriptional parts for scalable eukaryotic regulation using artificial DNA-binding domains. In addition, using recombinases, we have constructed synthetic-biological circuits that can store memory in genomic DNA and perform all two-input Boolean logic operations. The complexity of these circuits can be scaled by increasing the library of orthogonal recombinase devices.
We have also engineered synthetic gene circuits that implement complex analog computation. Using at most three transcription factors, we have created cells that compute wide-dynamic-range positive and negative logarithms, addition, subtraction, division, and power laws. We show that simple mathematical functions accurately capture the behavior of these analog circuits and can be composed together to achieve more complex functions. Moreover, we have generated detailed biochemical models which can explain the analog behavior of these computational devices.