Identification of metabolic units induced by environmental signals

Biological cells continually need to adapt the activity levels of metabolic functions to changes in their living environment. Although genome-wide transcriptional data have been gathered in a large variety of environmental conditions, the connections between the expression response to external changes and the induction or repression of specific metabolic functions have not been investigated at the genome scale.

We present here a correlation-based analysis for identifying the expression response of genes involved in metabolism to specific external signals, and apply it to analyze the transcriptional response of Saccharomyces cerevisiae to different stress conditions. We show that this approach leads to new insights about the specificity of the genomic response to given environmental changes, and allows us to identify genes that are particularly sensitive to a unique condition. We then integrate these signal-induced expression data with structural data of the yeast metabolic network and analyze the topological properties of the induced or repressed subnetworks. They reveal significant discrepancies from random networks, and in particular exhibit a high connectivity, allowing them to be mapped back to complete metabolic routes.