Genetic networks are often affected by stochastic noise, due to the low number of molecules taking part in certain reactions. In complex regulatory networks, noise in any one chemical species may induce noise in the rest of the system. In this paper, we analyse the sources of stochastic noise in the yeast galactose utilization pathway at the level of the complete system, by using both computer simulations, and experimental comparisons between wild-type yeast and a modified strain. A computer model was first used to determine the sources of network noise, as well as mechanisms by which noise is controlled. Results from an estimation tool, confirmed by detailed stochastic simulations, show that noise is caused primarily by fluctuations in mRNA concentrations due to their stochastic creation and decay. The noise is controlled by feedback loops which regulate transcription of certain genes. To test the effect of the feedback loops on the rest of the system, a modified strain of yeast was prepared in which regulation of two key genes is eliminated. As predicted by the model, the mutant strain is induced by galactose in a manner similar to that for the wild-type, but with a higher degree of stochastic noise.