High-throughput genomics has revolutionized our understanding of molecular-based processes in a wide variety of organisms. However, our ability to scale such information to higher levels of organization will ultimately determine the role that systems biology plays in ecology. The Hierarchical Experimental Responses at Macromolecular to Ecosystem Scales (HERMES) project was initiated to provide an experimental framework where transcript profiling, enzyme activity, and metabolic profiling could be associated with physiological and ecological data. Our results suggest that traditional analytical approaches, which focus on differentially expressed genes, are inadequate to capture the long-term response of plants to an environmental perturbation. We used instead a unique systems biology approach, based on weighted gene co-expression network analysis to overcome the problems of traditional analysis and better understand how a single gene change can impact a complex biological pathway. Specifically, highly correlated gene expression profiles were grouped into individual networks, termed modules, and verified computationally with gene enrichment analysis and visualization software. Modules were correlated to metabolite, enzyme activity, and physiological data to assess the significance of each pathway to phenotypic information (e.g., biomass, flowering, and seed production). We were able to use the resulting gene co-expression networks to scale from gene to pathway, and then pathway to phenotype.  In addition, a genomic signature was generated from metabolite and transcription data using a weighted similarity score algorithm to classify phenotypic response.  The use of both techniques allowed for the accurate prediction of phenotype and provides a putative molecular pathway encoding the phenotypic response.  Our challenge now is to scale the observed phenotypic responses to the level of population, community, and ecosystem.