Calculation of circadian power fraction from single-cell RevVNP signals. (A) Representative RevVNP fluorescence traces obtained from single-cell tracking of cells cultured at high density, showing raw data (yellow) and low-pass filtered data (black). (B) Representative RevVNP fluorescence traces obtained from single-cell tracking of cells cultured at low density, showing raw data (orange) and low-pass filtered data (black). (C–F) Schematic representation of the data analysis pipeline. A dataset of single-cell tracks (C) is pre-processed by low-pass filtering, mean subtraction and division by SD (D), using Fast Fourier Transform average PSD of the dataset is calculated (E) and the peak circadian frequency f_c is identified. This frequency is then used as the center of the integration window (dark gray) on the single-cell PSDs to calculate the circadian power fraction (F). (G–J) Analysis of single-cell tracks at high (G and I) and low (H and J) density, using the continuous wavelet analysis software pyBOAT, with signal pre-processing by mean subtraction and normalization by standard deviation. (G and H) Wavelet spectra. (I and J) Ridge detected from wavelet spectra. (K–N) Analysis of single-cell tracks at high (K and M) and low (L and N) density, using signal pre-processing by pyBOAT’s signal detrending. (K and L) Wavelet spectra. (M and N) Ridge detected from wavelet spectra. Wavelet spectra show a decrease in power after ∼30 h due to the prevalence of high RevVNP signals at the beginning of single-cell traces.