Fluorescence Correlation Spectroscopy (FCS) in cells often suffers from artifacts
caused by bright aggregates or vesicles, depletion of fluorophores or bleaching of
a fluorescent background. The common practice of manually discarding distorted curves
is time consuming and subjective. Here we demonstrate the feasibility of automated
FCS data analysis with efficient rejection of corrupted parts of the signal. As test
systems we use a solution of fluorescent molecules, contaminated with bright fluorescent
beads, as well as cells expressing a fluorescent protein (ICA512-EGFP), which partitions
into bright secretory granules. This approach improves the accuracy of FCS measurements
in biological samples, extends its applicability to especially challenging systems
and greatly simplifies and accelerates the data analysis. (C) 2010 Optical Society
of America
