Practical considerations for accurate estimation of diffusion parameters from single-particle tracking in living cells.

Advances in fluorescence microscopy have enabled high-resolution tracking of individual biomolecules in living cells. However, accurate estimation of diffusion parameters from single-particle trajectories remains challenging due to static and dynamic localization errors inherent in these measurements. While previous studies have characterized how such errors affect mean-squared displacement (MSD) analysis, practical guidelines for minimizing them during data acquisition and correcting them during analysis are still lacking. Here, we combine theoretical modeling and simulations to evaluate how exposure time and sampling rate influence the accuracy of MSD-based inference under fractional Brownian motion (FBM), a canonical model of anomalous diffusion. We demonstrate that decoupling exposure and sampling times enables escape from the error-prone regime, thus improving inference accuracy, and that incorporating an offset in nonlinear MSD fitting substantially improves the estimation of the anomalous diffusion exponent. We validate this framework using trajectories of cytoplasmic particles in Escherichia coli, recovering consistent diffusion parameters across multiple datasets. We further prove that the framework extends beyond FBM to general cases of subdiffusion, thereby offering practical strategies to improve both experimental design and data analysis in single-particle tracking of live or synthetic systems.