This project will entail the development and validation of model kernels representing regulatory motifs found in two key intracellular biochemical and genetic networks. The first network regulates transcription via the cAMP-inducible transcription factor CREB and plays a key role in determining long-term memory.
The second network underlies circadian rhythms and plays a key role in modulating cellular responses at different times of day. The regulatory motifs to be modeled are ubiquitous and appear in intracellular networks mediating a wide variety of cellular processes. Moreover, the general dynamical properties of these motifs and the regulatory networks they compose are well conserved and can likely be extended to a variety of other intracellular networks.
The model kernels will provide a number of methods for simulating the
evolution of dynamic variables. For example, ordinary differential equations
(ODEs) will be developed to precisely track the concentration of molecular
species over time. SBML kernels and models will be developed
and used to test BioSPICE modules. A methodology of modeling complex models
with BioSPICE will be examined, creating hybrid model kernels that combine
two or more of the above methods for simulating a network motif (e.g.,
regulation of gene transcription by transcription factors present in small
copy numbers).
The cAMP-inducible transcription model kernels will be validated with experiments that utilize neurons and small neural networks in primary cell culture. These experiments are essential in two ways. First, they will provide feedback concerning the accuracy of the mathematical models. Through iterative cycles of experimentation and modeling, the mathematical models will be developed as increasingly accurate representations of the corresponding biological processes. Second, the experiments will help determine key parameters, such as time constants, that govern the response of the network to stimuli. Reporter genes, regulated by CREB, will be introduced into cultured neurons and the dynamics of gene expression following the application of chemical neurotransmitters will be studied.
The approach taken here is innovative in that these models will be developed by a multidisciplinary team with a unique combination of experience in experimental neurobiology, molecular and genetic modeling, and software development. Moreover, the group has access to on-site, state-of-the-art facilities for both experimental neurobiology and computational modeling.
Another innovative aspect of this project is that the foundation for an object-oriented approach to model kernel development will be established. By constructing networks from components that are based on model kernels, each of which is independently well characterized and validated, computational models of intracellular networks can benefit from the well-established software design principles of structured programming, encapsulation and modularity. The object-oriented principle of inheritance provides additional advantages. As elements of intracellular networks inevitably will have some properties in common, it would be wasteful to respecify these common properties. Using inheritance, the common features of similar intracellular elements can be borrowed from existing model kernels and the modeler need only specify those properties that are unique.
Inheritance also reduces the incidence of errors that may be caused by repeated duplication of the same properties. Also, this approach makes modeling accessible to a wider audience of investigators regardless of expertise. Investigators with limited mathematical expertise can benefit from mathematically complex model kernels that have been developed by others. This approach should also promote standardization within the modeling community since the library will contain reliable model kernels that have been validated by other investigators.
This approach is promising because it makes use of software engineering principles that were specifically developed to address the concerns of designing increasingly complex software. The modeling community will eventually face similar concerns as it attempts to model increasingly complex intracellular biochemical and genetic networks.