EM-based algorithms for single particle tracking of Ornstein-Uhlenbeck motion from sCMOS camera data

Area of Science:

• Biophysics
• Computational Biology
• Imaging Science

Background:

• Single particle tracking (SPT) is crucial for understanding biomolecular dynamics.
• Accurate localization and parameter estimation are key challenges in SPT.
• Scientific complementary metal-oxide semiconductor (sCMOS) cameras are increasingly used for biomolecular imaging.

Purpose of the Study:

• To apply Expectation Maximization (EM) based algorithms to SPT data acquired with sCMOS cameras.
• To compare the performance of Sequential Monte Carlo - EM (SMC-EM) and Unscented - EM (U-EM) methods.
• To investigate the impact of dataset size on estimation accuracy and computational efficiency.

Main Methods:

• Implementation of EM algorithms for localization and parameter estimation in SPT.
• Utilized Sequential Monte Carlo (SMC) with particle filtering and smoothing for general distributions.
• Employed Unscented Kalman Filter and Unscented Rauch-Tung Striebel Smoother (U-KF/U-RTS) for computational efficiency in the U-EM method.

Main Results:

• Both SMC-EM and U-EM methods were successfully applied to SPT data.
• The study analyzed the influence of the number of images on the final parameter estimates.
• Computational efficiency of the U-EM method was investigated and compared to SMC-EM.

Conclusions:

• EM-based algorithms offer a robust framework for SPT analysis with sCMOS data.
• The choice between SMC-EM and U-EM may depend on specific application requirements regarding accuracy and computational resources.
• Further investigation into dataset size effects provides insights for experimental design in SPT studies.