Biophysical Imaging

Fluorescence correlation spectroscopy (FCS) determines mobility and stoichiometry of fluorescent species through statistical analysis of the autocorrelation of the fluorescence intensity signal. This information complements standard biochemical assays by allowing such information to be measured noninvasively in living cells. Unfortunately, autocorrelation analysis assumes stationarity and ergodicity of signals, which are conditions that are not always possible to obtain in living cells. For example, fluorescence correlation measurements at the nuclear envelope of living cells encounter a slow volume undulation process, which is superimposed on the diffusive fluctuations arising from motion of the fluorescent tag. This situation requires very long data acquisition times to ensure proper sampling of the slow process, which is difficult to achieve in live cell measurements. It would be advantageous to conduct short measurements that only sample the diffusive process. However, standard autocorrelation analysis does not allow choosing short segments to consider only the fluctuation of interest, because it assumes a data segment that is very long compared to the slowest timescale of fluctuations in the system. Here we present an analytical fluorescence correlation method that accounts for finite data segment length and recovers unbiased mobility and stoichiometry estimates from short data segments. We demonstrate this method using simulations and experimental data taken at the nuclear envelope of a living cell. Finally, we explore the possibility of fitting multiple data segment lengths simultaneously using this method to increase the resolving power of fluorescence correlation analysis. This work has been supported by grants from the National Institutes of Health (R01 GM098550, GM064589, and GM124279).