Genomes to Life Contractor-Grantee Workshop III
February 6-9, 2005, Washington, D.C.
Bioinformatics, Modeling, and Computation
49
Hybrid Bacterial Cell Models: Linking Genomics to Physiological Response
Jordan C. Atlas1* (jca33@cornell.edu), Mariajose Castellanos1, Anjali Dhiman1,2, Bruce Church2, and Michael L. Shuler1
1Cornell University, Ithaca, NY and 2Gene Network Sciences, Ithaca, NY
A major challenge in the biological sciences is to relate cell physiology to genomic structure. Models that explicitly link genomic and proteomic data to physiology are necessary to take full advantage of bioinformatics. We have developed a whole cell hybrid model that captures the dynamics of a single celled chemoheterotrophic prokaryote. By hybrid model we refer to inserting a genomically/molecularly detailed sub-model into a coarse-grained model which is embedded in a representation of the cell’s environment. The initial step is to construct a coarse-grained model with lumped “pseudochemical species” (lumped components of similar chemical species). All subsystems of the coarse-grained model can be “de-lumped” into genomically complete, chemically distinct subsystems with corresponding genes and gene products. Using this coarse-grained host model structure we expect to quickly build a complete coarse-grained model of any given chemoheterotrophic bacteria using data from chemostat or other growth experiments. By combining molecularly detailed modules within the coarse grained host model, we capture not only the internal details of the dynamics of the molecular subsystem, but also can evaluate that mechanism within the context of a whole cell and its environment. The whole cell modeling approach presented here is being augmented by statistical mechanics methods for parameter estimation that allow us to rapidly develop parameter sets for new modules as they are added.
This framework has been applied to create Cornell’s Minimal Cell Model (MCM). The MCM is a theoretical construction that attempts to develop our understanding of the relationship between cellular function and genetics using a “bottom up” approach; the necessary model functions are selected by rationally deciding what machinery a cell needs to live and reproduce. A “minimal cell” is a hypothetical free living cell possessing the functions required for sustained growth and reproduction in a maximally supportive culture environment. The MCM simulates the growth of a minimal cell. Ultimately, we aim to model the complete functionality of a minimal cell. The Shuler group has demonstrated the “modularity” of hybrid models by constructing a genomically and chemically detailed model of nucleotide metabolism within the MCM (PNAS v. 101(17), pp. 6681-6686). The current work focuses on incorporating amino acid supplementation into the coarse grained model. Another system of interest is Shewanella oneidensis, which has the potential to help remove metal pollutants from the environment. We believe that these techniques will ultimately allow us to build a model for Shewanella that creates a connection from the organism’s genomics, to its molecular functions, to the whole cell, and to the environment.
* Presenting author
