METABOLIC NETWORKS FCModeler and PathBinder for Network Modeling and Creation

1. INTRODUCTION The field of systems biology in living organisms is emerging as a consequence of publicly-available genomic, transcriptomics, proteomics, and metabolomics datasets. These data give us the hope of understanding the molecular function of the organism, and being able to predict the consequences to the entire system of a perturbation in the environment, or a change in expression of a single gene. In order to understand the significance of this data, the functional relationships between the genes, proteins, and metabolites must be put into context. This chapter describes an iterative approach to exploring the interconnections between biomolecules that shape form and function in living organisms. We focus on the model plant system, Arabidopsis. The systems biology approach itself can be used as a prototype for exploration of networks in any species. The biologist's information about the function of each RNA and protein is limited. Currently, about 50% of Arabidopsis genes are annotated in databases (e.g., TAIR 1 or TIGR (www.tigr.org)). In part because the process of evolution results in families of genes with similar sequences and related 2 Chapter #17 functions, much of the available annotation is not precise, and some annotation is inaccurate. Even more limited is our understanding of the interactions between these biomolecules. To help bridge this gap, metabolic networks are being assembled for Arabidopsis (e.g., AraCyc, KEGG). To date, these contain many derived pathways based on other organisms; consequently, they have errors and do not capture the subtleties of the Arabidopsis (or even plant) biochemistry and molecular biology that are necessary for research. Considerable high-quality information is buried in the literature. A given pathway is known predominantly to those researchers working in the area. Such a pathway is not easily generated by curators whom are not experts in the particular field. This information is not rapidly accessible to a biologist examining large and diverse datasets and investigating changing patterns of gene expression over multiple pathways in which she/he may have little expertise. Furthermore, the interconnections between the multiple complex pathways of a eukaryotic organism cannot be envisioned without computational aid. To assist biologists in drawing connections between genes, proteins and metabolites, cumulative knowledge of the known and hypothesized metabolic and regulatory interactions of Arabidopsis must be supported by advanced computing tools integrated with the body of existing knowledge. Metabolic Pathway Databases There are a few major database projects designed to capture pathways: What Is That? …