A Hidden Markov Model for Detecting Confinement in Single-Particle Tracking Trajectories

Area of Science:

• Biophysics
• Membrane Biophysics
• Statistical Mechanics

Background:

• Single-particle tracking (SPT) generates high-resolution trajectories for studying particle dynamics in membranes.
• Existing statistical methods partition trajectories into diffusion states but lack detailed confinement analysis.
• Understanding diffusion states and confinement is crucial for interpreting molecular behavior in biological membranes.

Purpose of the Study:

• To develop a hidden Markov model for analyzing particle confinement and free diffusion states in SPT data.
• To implement a Markov chain Monte Carlo algorithm for automated trajectory partitioning.
• To characterize the heterogeneity of confinement events in biological membranes.

Main Methods:

• Developed a hidden Markov model with phases of free diffusion and harmonic potential confinement.
• Employed a Markov chain Monte Carlo algorithm for model fitting and trajectory partitioning.
• Applied the model to interferometric scattering microscopy data of GM1 lipids in model membranes.

Main Results:

• Successfully automated the partitioning of SPT trajectories into free diffusion and confinement phases.
• Revealed heterogeneity in the lifetime, shape, and size of confinement events.
• Demonstrated that confinement heterogeneity arises from nanoparticle characteristics and the binding-site environment.
• Observed that confinement size and shape are conserved within individual trajectories.

Conclusions:

• The developed confinement model accurately detects and characterizes diffusion states in SPT data.
• Heterogeneity in confinement events provides insights into molecular interactions and membrane organization.
• The model has broad applicability for studying various biological phenomena like receptor clustering and lipid rafts.
• Nanoparticle tag characteristics influence the resolution limits of SPT, suggesting deconvolution methods for homogeneous tags.