In late September, the Defense Advanced Research Projects Agency awarded several contracts and grants under its Bio-Computation program, a five-year effort with about $65 million in program funding to support R&D in both DNA computing and computational modeling of cellular systems of interest to the US Department of Defense.
Program director Sri Kumar told BioInform that 12 grants were awarded under the latter half of the program to support the development of BioSPICE (Simulation Program for Intra-Cell Evaluation). BioSPICE will be an open source software package based upon a model library or kernel of relevant biochemical processes, gene-protein interactions, and analytical tools, with models validated by cell experiments. The goal is to integrate this kernel into a user-friendly simulation environment with links to related databases.
If you can effectively model cellular processes then you can predict behavior and the response to external events and also design controls to find targets that can be used for intervention and therapeutic design, Kumar said.
The general goal of the project is to significantly reduce the time and cost associated with designing drugs and identifying targets, Kumar said, or as some describe the long-term vision, bug to drug in 24 hours. However, with the threat of biological warfare looming in the wings, Kumar said the heightened need for these tools is clear.
Within the DOD theres a lot of interest right now in trying to analyze in silico the pathogenicity of bugs, he said.
Solicitations for the program went out in May, Kumar said, and the awards were not directly related to the September 11 terrorist attacks.
The tools are expected to be used on a number of levels. Due to its open source nature, BioSPICE will be available for the entire biological community to use once completed. Yearly releases are planned with the first release expected in about a year. In addition, Kumar said, other DARPA programs, such as the biosensors program and the unconventional pathogen program, will be able to use the models and tools, and DARPA is working with the NSF and NIH to make them available for use and development to researchers from these agencies as well.
Last week, BioInform spoke to several of the project leaders working on different aspects of the project:
SeekING Network Motifs
One of the larger projects is led by Adam Arkin from the University of California, Berkeley, and Lawrence Livermore National Laboratory to study models of sporulation and germination in several organisms, primarily Bacillus subtilis.
While B. subtilis is not a pathogen, Arkin said the organism is of interest to DARPA because the sporulation and germination pathways are well conserved among a large class of Bacilli, including B. thuringiensis (a pesticide), B. cereus (an enteropath), and B. anthracis (the bacteria that causes anthrax).
Arkin, who began developing BioSPICE with DARPA funding in 1995, said he is also working very closely with the SRI team that is integrating the various BioSPICE components. One of the key features he envisions for the software is an ability to perform comparative network analysis. Just like theres comparative genomics that looks at sequence motifs, there are network motifs, Arkin said. Were interested in how these network motifs create similar but different strategies in these organisms and being able to deduce those new strategies by using information from a known model organism.
In particular, the Arkin group is studying modes of failure in the sporulation process. Theres a five percent failure rate in the wild type and we want to know how that failure rate arises how do we cause or prevent sporulation and germination?
While the team has existing models of pieces of the pathway for sporulation initiation in place, Arkin said a fully rendered model is about three years away.
Cell Growth and Division
John Tyson of Virginia Polytechnic Institute is heading up a consortium of experimentalists, modelers, and computer scientists from four institutions: Virginia Tech, the Technical University of Budapest, Rockefeller University, and the University of Kentucky Medical School. Tyson said the group is aiming to provide a prime example of the usefulness of this approach to understanding a fundamental problem in cell biology: the regulation of cell growth and division.
The modeling group is focusing on the molecular regulatory system underlying growth and division in yeast cells and frog embryos. Meanwhile, molecular biologists will test predictions of the models in genetically modified yeast cells, in frog egg extracts, and in intact frog embryos. Biologists will also collect quantitative information about molecular components during the cell cycle in order to provide data to improve the predictive power of the models.
Finally, computer scientists from Virginia Tech will contribute software and expertise in the areas of model building, comparison of simulations to experiment, parameter estimation, nonlinear dynamics, and bifurcation theory.
ModelING Toggle Switches
Jim Collins, professor of biomedical engineering and co-director of the Center for BioDynamics at Boston University, applies concepts from nonlinear dynamics to problems in biology, trying to characterize, improve or mimic biological function. As part of the DARPA project, he will attempt to develop a simulation technique that spans Monte Carlo technique and a rate equations approach. This would be quite efficient in that it could switch on the fly between these two representations, Collins said.
Collins will apply the simulation technique to simple gene regulatory networks such as a single gene system with a positive feedback or a genetic toggle switch system where two genes are trying to switch each other off and will eventually test his models experimentally by building these networks in bacteria or yeast.
In a recent study on a single-gene feedback system in E. coli that is soon to be published, the Collins team found that their model provided accurate predictions of the dynamics of the real system. The researchers then modified the model to account for the intrinsic noise of natural systems by incorporating stochastic fluctuations. Its that interplay between the theory and the model and the experiments that lies at the heart of our proposal, Collins said.
One long-term defense-related application of a genetic toggle switch might be bacterial sensors that can detect biological pathogens. According to Collins, these could be used to monitor water supplies and could be applied in remote settings like the desert.
Studying Mammalian Cells
Bud Mishra, professor of computer science and mathematics at the Courant Institute of Mathematical Sciences at New York University and a faculty member of Cold Spring Harbor Laboratory, is collaborating with CSHL cancer biologist Michael Wigler.
Mishras group is developing informatics, simulation, and reasoning tools for use in modeling biological experiments in mammalian cell lines. The results could be used to devise new wet experiments to test hypotheses and revise models.
Mishra and his students are using new quantitative sequence analysis tools called Valis to model DNA-repair and -recombination processes. Based on Wiglers experimental results, Mishra is also modeling changes in gene transcription as a function of co-cultivation of two cell types, such as two different cancer cell lines. The purpose is to identify receptor-ligand pairs that initiate the signaling processes that lead to the changes in transcription. Host-pathogen interactions could be studied with that, Mishra noted a possible defense-related application. Further down the line, his tools could also be used to model circadian rhythms, sleep-wake cycles, and other processes of interest to the DoD.
BT and JK