Digital signal processing has been a cornerstone of the modern communications and electronics revolution, transforming application areas such as wired and wireless communications, storage, and biomedical signal processing. This project will study digital signal processing in an entirely new domain: molecular systems. In contrast to electronic systems, where signals are represented by time-varying voltage values, in molecular systems signals are represented by time-varying concentrations of different molecular types, such as proteins, RNA and DNA.

This project will develop, implement, and evaluate molecular-level designs for a variety of digital signal processing operations such as filtering, equalization, and noise cancellation. These operations will be synchronized by clock signals, created through sustained chemical oscillations. Memory will be created by transferring signals between different molecular types in alternating phases of the clock. The key idea underpinning this research is that the computation should be essentially rate-independent: it should only depend on coarse categories for the rates of the chemical reactions (e.g., ?fast? vs. ?slow?). It should not matter how fast any ?fast? reaction is ? only that ?fast? reactions are fast relative to ?slow? reactions. Designs with this property can be mapped to different chemical substrates. They compute accurately in spite of variations in environmental conditions such as temperature.

The impetus for this work is not computation per se; chemical systems will never be useful for number crunching. Rather the field of molecular computing aims for the design of custom, embedded biological ?sensors? and ?controllers? ? viruses and bacteria that are engineered to perform useful tasks in situ, such as cancer detection and drug therapy. As an experimental chassis, this project will map designs for digital signal processing operations to chemical reactions involving DNA strands. These designs will be evaluated with computer simulations of the chemical kinetics.

Techniques for analyzing the dynamics of biological systems are well established. However, synthesizing computation with such mechanisms requires new techniques ? and an entirely new mindset. The digital circuit design community has unique expertise that can be brought to bear on the challenging design problems encountered in synthetic biology. Applications in biology, in turn, offer a wealth of interesting problems in algorithmic development. With its cross-disciplinary emphasis, this project will bring new perspectives to both fields.

If successful, the proposed research will transform disciplines such as genetic engineering of drug-delivery systems. Currently, a costly, ineffective ad-hoc approach prevails. With robust and rate-independent techniques for implementing operations such as digital signal processing, much more effective systems will be developed. An important goal of the project is to communicate the impetus for interdisciplinary research to a wide audience. A new course will be developed, titled "Circuits, Computation, and Biology" offered jointly through the Electrical Engineering Department and the Biomedical Informatics and Computational Biology Program at the University of Minnesota. Building upon current recruitment efforts that have brought in female students, students from underrepresented groups will be recruited into the project.

Project Report

Digital signal processing continues to play a major role in our daily lives. Whenever a speech or image is transmitted by a wired or wireless medium or is uploaded or downloaded from the internet, signal processing is used in the form of speech coding or image coding, modulation and transmission. Apart from communications, signal processing plays a major role in biomedical signal analysis as well as implantable and wearable healthcare devices such as pacemakers. This project was focussed on synthesis of discree-time signal processing systems using molecular reactions in general and DNA strands displacement reactions in particular. The objectives include activation or inhibition of certain pathways, protein monitoring including cancer cell growth monitoring and drug delivery. The input signals are sampled by a clock sythesized by DNA transfer reactions. The sampling of the signals can be achieved in one of three ways: a synchronous approach using a two-phase non-overlapping clock, a three-phase globally-synchronous locally-asynchronous scheme, and a fully asynchronous scheme. Tradeoffs in these implementations were investigated. Various signal processing functions such as finite-impulse response (FIR), and infinite-impulse response (IIR) digital filters and fast Fourier transforms (FFTs) were synthesized using DNA. These systems were simulated using computers. The challenge of long sampling periods in the order of one day is a major bottleneck of the proposed method. Current work is being directed towards overcoming this bottleneck. Sequential systems and random logic gates were also simulated using DNA. The proposed synthesis methods require less number of DNA reactions than known approaches.

Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Minnesota Twin Cities
United States
Zip Code