On April 14, 2004, the National Academies Keck Futures Initiative announced the 11 recipients of the first Futures Grants, each in the amount of $75,000, to support interdisciplinary research on signaling. The research projects that secured funding represent a wide range of approaches to exploring cell signaling systems, which was the subject of the first Futures conference, held in Irvine, Calif.
These competitive grants provide seed funding to researchers, who participated the Signals, Decisions, and Meaning in Biology, Chemistry, Physics, and Engineering conference, to enable them to pursue important new ideas and connections stimulated by the conference. The grants aim to fill a critical missing link between bold new ideas and major federal funding programs, which do not currently provide grants in areas that are considered risky or unusual. The seed grants allow researchers to start developing a line of inquiry by recruiting students and postdoctoral fellows, purchasing equipment, and acquiring preliminary data all of which can position the researchers to compete for larger awards from other public and private sources.
The Signaling award recipients and their grant research topics are (affiliations noted are at the time the grants were awarded):
LEE BARDWELL University of California, Irvine
Intracellular Signaling Specificity: Theoretical and Computational Explorations
Bardwell will use his Keck Futures grant for an interdisciplinary computational and mathematical investigation of the mechanisms that promote specificity in cell signaling. This will complement ongoing experimental work in his laboratory. Cells react to a wide variety of stimuli, such as chemical messages sent by other cells. Each different stimulus triggers a distinct cellular response; yet cells use a limited number of component proteins to transmit these numerous signals. Bardwell will investigate how modular, adhesive protein-protein interactions and feedback circuits collaborate to provide specificity in such situations. Finally, Bardwell will explore aspects of the evolution of signaling pathways.
TOBY BERGER Cornell University, Ithaca, N.Y.
Directed Information in Neural Networks With Feedback
Berger proposes to extend in several directions the discrete-time mathematical model of energy- efficient information signaling within and among regions of sensory cortex. He will endeavor to replace the current unrealistic assumption that all post-synaptic potentials (PSP) are quenched between successive time slots with a neuroscientifically more accurate one to the effect that a neuron’s PSP gets fully quenched only if it just generated an action potential; otherwise, it suffers only a leakage decay. Also, we shall strive to incorporate the observed phenomenon that the probability of a synaptic failure is a function both of whether or not a failure occurred at the synapse in question the previous time an afferent spike arrived there, and of how much time has elapsed since a previous arrival.
Poster Presentation Abstract for Neuroscience 2004
Encoding of excitation via dynamic thresholding.
T. BERGER. School of Electrical. and Computer Engineering, Cornell University, and W. B. LEVY, Department of Neurosurgery, University of Virginia Medical School, Charlottesville, VA
We analyze how a visual system neuron whose spiking threshold continuously descends for tens of ms after each of its action potentials could enable the neuron to report reliably and rapidly about variations in the intensity of its afferent bombardment via the timing between its successive efferent spikes. Whereas a neuron whose spiking threshold remains constant after a brief absolute refractory period is known to exhibit near-unity coefficients of variation of interspike interval durations over almost the entire range of interesting bombardment intensities, dynamic threshold descent enables useful timing codes at all levels of excitation. The mathematical theories of Poisson approximation, filtered non-homogeneous Poisson processes, and estimation of the time-of-arrival of a known waveform in additive noise together explain how a steadily descending spiking threshold can enable a simple, natural, precise, broadband timing code mechanism that could readily be effected by the biology of sensory neurons. .This work was supported in part by a National Academies Keck Futures Initiative Grant to TB and by MH63855 to WBL.
JAY DUNLAP Dartmouth Medical School, Hanover, N. H.
Existence of Autocatalysis Based on FRQ Protein Proteolysis
The fundamental question of circadian rhythms is, what is the mechanism that drives a biochemical oscillator of about 24 h satisfying the following properties: endogenous rhythmicity in constant conditions; entrainment to external cues (i.e., light and temperature); resetting by environmental cues; and temperature and nutritional compensation. Dunlap's project proposes to test an alternative model where autocatalysis plays a role equally as important as the existing negative feedback loop in generating a robust biochemical oscillator. This model incorporates a hysteresis driven by the positive feedback loop, along with a stable limit cycle due to both feedback loops. Dunlap and his colleagues will test this model in the genetically, biochemically, and molecularly tractable Neurospora system in which FRQ acts as the essential negative element in the circadian feedback loop.
JAMES FERRELL Stanford University, Stanford, Calif.
Bistability as a Mechanism for Forming Domains Without Walls
Signaling systems that include a positive feedback loop, or a double-negative feedback loop, can, under the proper circumstances, exhibit bistable behavior. A bistable system is one that can be made to toggle between two alternative, stable steady-states. Bistability can serve as the basis for all-or-none biochemical responses, for sustained “memories” of transitory stimuli, and for the temporal segregation of successive waves of a multiphase or oscillatory response. Farrell proposes to examine whether bistability may also allow the maintenance of discrete spatial domains within a continuously diffusing cytoplasm. He proposes to take experimental, computational and mathematical approaches to this question.
HEIDI HAMM AND EMMANUELE DIBENEDETTO Vanderbilt University, Nashville, Tennessee
Mathematical Models in Signaling Systems
The goal of this project is to understand signal integration of multiple G protein pathways. The thrombin receptor, PAR1, is known to mediate its complex actions in the endothelium via its ability to couple to multiple G proteins: Gi, Gq and G12/13. To assist in the direction of their research and to encapsulate their findings. Hamm and DiBenedetto will introduce a mathematical model which incorporates the spatio-temporal aspects of signaling. By using a modular modeling approach, we will iteratively build toward our goal of producing a validated computational model of the complex signaling processes used in cellular communication.
HOD LIPSON Cornell University, Ithaca, N.Y.
URI ALON Weizmann Institute of Science, Rehovot, Israel
Evolutionary Computation Methods for Inference of Signal Transduction Networks
Lipson and Alon aim to develop and apply new computational methods for inference of signaling networks within the cell using automated physical experimentation and hypothesis generation. The approach is based on new engineering topological system identification methods combined with a unique automated experimental system capable of accurately controlling and measuring nutrients in cultured systems over time. This approach actively studies the cellular system by generating hypotheses, and then planning and executing experiments to verify, refute, and elucidate these hypotheses in the most efficient way. The research will first be applied to the well-studied lactose and arabinose gene regulation systems in E. Coli, paving the path towards automated reconstruction of more complex, unknown systems in the future.
Conference Paper: Genetic and Evolutionary Computation Conference 2004
Josh C. Bongard and Hod Lipson, (2004) "Automating Genetic Network Inference Using a Very Low Sampling Estimation-Verification Evolutionary Algorithm", Genetic and Evolutionary Computation Conference, (GECCO '04), pp. 333-345 http://www.mae.cornell.edu/ccsl/papers/GECCO04_Bongard.pdf
TERENCE HWA University of California, San Diego
Specificity, Cross Talk, and Evolvability in Two-Component Signaling
Hwa proposes experiments to study molecular specificity vital to the fidelity of signal transduction in cells. He will focus on the two-component signaling system in E. coli, and take an evolutionary approach to breed signaling proteins with different degrees of interaction specificity via a duplication/divergence process. Bioinformatic and computational studies will be performed on the evolved protein sequences to determine how different degrees of specificity and cross talk can be achieved molecularly, and to assess how evolvable two-component systems can be in the development of novel signaling pathways.
SHARAD RAMANATHAN Bell Laboratories, Lucent Technologies, Murray Hill, N.J.
A Study of MAP Kinase Pathways in Yeast
A haploid yeast Saccharomyces cerevisiae cell decides on one of several developmental fates depending on its environmental conditions. A long-standing puzzle has been how the different MAP Kinase pathways that respond to pheromones, glucose starvation, or osmolar pressure in the environment, manage to retain their fidelity to the respective inputs despite sharing some components. Ramanathan proposes to understand the role of the MAP Kinase module in the decision-making process of the cell through a combination of experimental and analytical approaches.
RAMA RANGANATHAN University of Texas Southwestern Medical Center and Howard Hughes Medical Institute, Dallas
Robustness and Evolvability in Proteins
Evolution builds proteins that, like finely tuned machines, are beautifully suited for mediating complex functions. Nevertheless, they somehow remain remarkably tolerant to random mutation, and are evolvable; that is, they can change to adapt to altered functional pressures. In this grant, we propose to develop new methods for studying the origin of these properties in proteins, and then plan to test hypotheses through building artificial proteins. The ultimate goal is to understand the design principles in evolution that account for high performance function while maintaining stability to the process that created them (random variation) and adaptability to manage in an unpredictable and ever-changing world.
PAMELA SILVER Harvard Medical School and the Dana-Farber Cancer Institute, Boston
Biological Pattern Formation via the Design and Construction of Integrated Cell Signaling Systems
Silver and Drew Endy of the Massachusetts Institutes of Technology will study biological pattern formation via the design and construction of integrated cell signaling systems. The goal of this research is to apply past lessons from design of computer chips to self-replicating molecular systems, such as biological cells. This research is at the forefront of a new discipline termed Synthetic Biology, which focuses on large-scale synthesis of biological systems. Recent advances in DNA synthesis and sequencing make it possible to manufacture DNA at the genome scale. However, there is almost no accompanying set of design tools that allows researchers us to make good use of DNA synthesis as a fabrication resource. This research will allow the creation of a bioengineering grammar for large-scale DNA synthesis and offers an opportunity for synergy between cell biology and engineering.
MICHAEL SIMPSON University of Tennessee and Oak Ridge National Laboratory, Oak Ridge, Tenn.
The Processing of Stochastic Fluctuations in Gene Circuits and Networks
The goal of Simpson's project is to advance the understanding of stochastic fluctuation processing in biological systems. He focuses on the development of hardware acceleration of stochastic simulation with a long term vision of continued simulation, speed increases as an automatic consequence of the Moore’s law advance of the semiconductor industry; and the extension of the frequency domain analysis approach to address fundamental issues in the processing of stochastic fluctuations within gene circuits and networks, including stochastic modulation of nonlinear biochemical processes, stochastically valid approximate relationships for gene expression control, and the processing of noise in cell-cell communication systems.