Application limits and data correction in number of molecules and brightness analysis

Number of molecules and Brightness (N&B) has been proposed for measuring the molecular brightness and number of fluorophores in time-sequence of images, in live cells. If the fluorescently tagged-proteins are mobile in the illumination volume, the stoichiometry of their oligomers can be derived from the increase of the brightness of the fluorescent dyes due to clustering. We examine aspects concerning extra-fluctuation effects induced by cell shifts and photobleaching, which yield large overestimates of the clusters size and sub-unit counts. We develop an offline corrective approach consisting in frame re-alignment and boxcar filtering for recovering precision of the analysis. Using simulations we derive general criteria for approaching this analysis, and assess the application limits of the corrective procedure. We tested the approach in extreme experimental conditions (few pixels, large extra-variance perturbations), in which we analyzed the minimal increases of brightness as that expected between a monomeric and dimeric GPI-mEGFP constructs. We show how most of the perturbing effects can be abolished, and obtain the correct the brightness of GPI-mEGFP monomers and dimers.