Period-phase map: two-dimensional selection of circadian rhythm-related genes

Many genes related to the circadian rhythm, especially those involved in phase shifts induced by different environmental stimuli, still remain enigmatic. In this study, the authors monitored the expression of rat genes measured with multiple phase-resetting stimuli, and developed a technique to extract the candidate genes for the changes in circadian rhythm by the stimuli, from microarray data. First, the spectra for the time series of gene expression were estimated by fast Fourier transform, and then two fitting methods, the random period fitting method and the conditional curve fitting method, using the estimated periods as the initial values, were applied to the control and the stimulated expression data to estimate the periods and the phases. Finally, by comparing the two sets of periods and phases, the period change and the phase shift by stimuli were estimated to extract the candidate genes related to the master clock, by mapping the period change and the phase shift on a two-dimensional space, a period-phase map (PPM). As an indirect validation of the genes selected by our method, the significant enrichment of extracted gene clusters on the PPM was further evaluated, in terms of biological function. As a result, the gene clusters related to photoreceptors and neural regulation emerged on the PPM, thus implying the relationships in the stimulus response of the master clock that resides in the brain at the intersection of the optic nerves. Thus, the present approach is a feasible means to explore the oscillatory genes related to stimulus responses.