Karl Runge, Ph.D.

Senior Research Scientist
Gene Network Sciences


Implementation of the GNS Reverse Engineering Process at CCNI

Gene Network Sciences' (GNS) Reverse Engineering/Forward Simulation (REFS) system is a computationally demanding analysis process that enables the inference of biological models from experimental data.

In this talk we describe the implementation of two key parts of the GNS REFS process: 1) Model Enumeration and scoring, and 2) Global Optimization and Monte Carlo sampling of biological models. Both of these REFS components have been implemented to use the remote resources available at the Computational Center for Nanotechnology Innovations (CCNI). The BlueGene/L system and Opteron Linux cluster resources at CCNI are utilized, and have enabled GNS to achieve a 40-fold increase in computational throughput in a commercial, production environment. Additionally, the above speedup coupled with the increased memory resources at CCNI have allowed GNS to work on problems that are 10-20 times larger than could be performed previously.

Along with an overview of the algorithms used in REFS, we also discuss the integration of the CCNI resources into the GNS REFS platform we have implemented. This integration enables application scientists at GNS to seamlessly tap into the computational resources at CCNI by simply setting a queue parameter indicating whether the computational task should be run remotely at CCNI or on local resources.


Karl Runge received his Ph.D. in Theoretical Condensed Matter Physics from Cornell University in 1988. He has held positions in scientific computing and software programming. He is currently employed by Gene Network Sciences as a Senior Research Scientist.

updated: 2008-09-16