Gene expression density profiles characterize modes of genomic regulation: theory and experiment.

Our study addresses modes of genomic regulation and their characterization using the distribution of expression values. A simple model of transcriptional regulation is introduced to characterize the response of the global expression pattern to the changing properties of basal regulatory building blocks. Random genomes are generated which express and bind transcription factors according to the appearance of short motifs of coding and binding sequences. Regulation of transcriptional activity is described using a thermodynamic model. Our model predicts single-peaked distributions of expression values the flanks of which decay according to power laws. The characteristic exponent is inversely related to the product of the connectivity of the network times the regulatory strength of bound transcription factors. Such 'expression spectra' were calculated and analyzed for different model genomes. Information on structural properties and on the interactions of regulatory elements is used to build up a framework of basic characteristics of expression spectra. We analyze examples addressing different biological issues. Peak position and width of the experimental expression spectra vary with the biological context. We demonstrate that the study of the global expression pattern provides valuable information about transcriptional regulation which complements conventional searches for differentially expressed single genes.