Myers Research Group

University of Utah



People

Research

Tools

Publications

Book

Home Page

A Principled Mapping of Regulatory Networks to Asynchronous Circuit Models for Stochastic Analysis

SUPPORTED BY

National Science Foundation Quantitative Systems Biology Program

Through the use of microarrays and other new technologies, we are beginning to get vast amounts of data on how genes interact to perform complex biological functions. In order to reason about genetic systems, a systems biology perspective must be taken in which new models and efficient analysis methods must be developed. Electrical engineers have vast experience modeling and analyzing electronic circuits and systems. McAdams and Shapiro in their 1995 Science paper took an electronic circuit view of a genetic network with encouraging results. Therefore, as in the sequencing of the human genome, collaborations between engineers and systems biologists may be extremely beneficial to the success of functional genomics.

In collaboration with Professor Adam Arkin's Laboratory for Dynamical Genomics at UC Berkeley, we have been working on modeling the Phage lambda virus using a stochastic asynchronous circuit model. A stochastic model appears to be essential as the survival strategy taken by this virus has a random component which may be key in the evolutionary survival of this and other species. Dr. Arkin's original model based on the chemical master equation and Monte Carlo simulation required substantial runtime on a supercomputer while our new stochastic asynchronous circuit model produces comparable results in under a minute on a PC. Our results also show good agreement with prior experimental data.

The Phage lambda case study has led to the development of an abstraction methodology from reactions with kinetic rates and critical concentrations to a stochastic asynchronous circuit model. After this abstraction, efficient Markov chain analysis methods can be applied to reason about the systems behavior. The first major goal of this project is to apply this methodology to other systems that exhibit stochastic behavior. Candidate systems include the the E. Coli Fim system and B. Subtilis stress response network. Development of the models, analysis of results, and coordination of laboratory experiments to validate predictions is being carried out through a collaboration with Dr. Arkin's lab. Our primary anticipated research result is a complete methodology for the efficient analysis of genetic regulatory networks and its demonstration on several example systems.

While this research represents a significant change in direction for us, the differences in the problem domains are perhaps not as great as appear on the surface. We have developed efficient methods for the synthesis, analysis, and verification of timed asynchronous circuits. We have also developed analysis methods for analog error control decoder circuits. Genetic networks are certainly asynchronous systems (there is no global clock), and at times they must also be considered analog in that the real concentration levels of species must be taken into account. Therefore, there is a lot of promise to adapting methods for the modeling and analysis of asynchronous and analog circuits to those of genetic regulatory networks. In other words, a major goal of this research project is to adapt our prior research results to a new problem domain, that of genetic networks.

Long term societal impact of this research is significant. The efficient analysis of biological systems in silico has the promise of helping our understanding of the causes of disease and our development of drugs to treat them.

Faculty:

Students:

Publications:

Tools:

Courses: