Team uses NCSA resources for gene networks simulation code

Story posted February 1, 2006

With the ability to command cells to make specific proteins, researchers will be able to generate specialty compounds such as therapeutic drugs, insulin, or sensors that detect threatening biological vectors like anthrax. To exercise that command, they must understand gene regulation (when and how DNA produces a given protein) and engineer gene networks that produce those proteins. Yiannis Kaznessis and his team at the University of Minnesota use a variety of NCSA platforms to predict the dynamic behavior of gene networks and to develop the code to do so.

Recently they released their code, Hy3S, to the biology community. The multiscale algorithm integrates stochastic-discrete, stochastic-continuous, and deterministic-continuous models of biomolecular interactions. This powerful combination has already allowed them to develop principles for genetically engineering bistable switch networks and oscillatory networks. Results have been published in the Journal of Chemical Physics and Biophysical Journal. These principles are now being realized experimentally by others in the field.

Besides the sheer number of computing cycles he can secure on NCSA machines, "NCSA resources are attractive because of the wealth of available platforms that result in no down time for our runs," according to Kaznessis. "Storage capacity and the piece of mind that our results are carefully stored and easily retrieved -- through the center's UniTree system is another reason."

This research is supported by the National Science Foundation's Bioengineering and Environmental Systems Division through grant BES-0425882.