Minimization of transcriptional temporal noise and scale invariance in the yeast genome

The analysis of transcriptional temporal noise could be an interesting means to study gene expression dynamics and stochasticity in eukaryotes. To study the statistical distributions of temporal noise in the eukaryotic model system Saccharomyces cerevisiae, we analyzed microarray data corresponding to one cell cycle for 6200 genes. We found that the temporal noise follows a lognormal distribution with scale invariance at the genome, chromosomal and sub-chromosomal levels. Correlation of temporal noise with the codon adaptation index suggests that at least 70% of all protein-coding genes are a noise minimization core of the genome. Accordingly, a mathematical model of individual gene expression dynamics was proposed, using an operator theoretical approach, which reveals strict conditions for noise variability and a possible global noise minimization/optimization strategy at the genome level. Our model and data show that minimal noise does not correspond to genes obeying a strictly deterministic dynamics. The natural strategy of minimization consists in equating the mean of the absolute value of the relative variation of the expression level (alpha) with noise (eta). We hypothesize that the temporal noise pattern is an emergent property of the genome and shows how the dynamics of gene expression could be related to chromosomal organization.