The coefficient of variation by image correlation spectroscopy (CV-ICS) quantifies chromatin compaction from images with low photon counts

Lower excitation light levels reduce phototoxicity in confocal and superresolution microscopy but generate images with lower photon counts. A well-established parameter that quantifies chromatin compaction from fluorescence images of labeled nuclei is the coefficient of variation (CV), which is defined as the ratio between the standard deviation and the average of the pixel intensity values. Here, we show that when imaging is performed at low photon counts the standard deviation of the intensity is dominated by the shot noise fluctuations and the CV is not representative of the chromatin compaction state. To overcome this issue, we introduce an alternative calculation of CV based on image correlation spectroscopy (CV-ICS), where the CV-ICS parameter is extracted from the spatial autocorrelation function of the image. We demonstrate the use of CV-ICS on fast confocal (resonant) imaging of live HeLa cells labeled with Hoechst 33342. First, we show that the conventional CV calculation deviates already by about 50% from the correct value when the maximum photon counts per pixel are below 40. In contrast, the CV-ICS calculation is unaffected by the presence of shot noise. Next, we use imaging at low photon counts and CV-ICS to follow changes in the chromatin compaction state after switching the cells to a medium of higher (hyper-) or lower (hypo-) osmolarity. Finally, we use CV-ICS to investigate the time course of chromatin compaction after induction of laser-induced DNA damage. This work shows that CV-ICS can be a robust parameter for quantifying chromatin compaction in all imaging applications with a limited photon budget, including live cell imaging and superresolution microscopy.