Connecting the dots: probabilistic mapping of biomolecular paths in fluorescence microscopy

The concerted behavior of biomolecules bring life to cells. Understanding this behavior requires detecting and tracking these molecules over time.

Due to the nanometer precision in super-resolution microscopy, the main obstacle for use of fluorescence microscopy for this kind of dynamics is no longer inadequate spatial resolution.

Rather, the challenge is to reliably turn stacks of images of fluorescent biomolecules into trajectories including associated biophysical analysis.

This proposal shifts the perspective on the analysis used in fluorescence microscopy of biomolecules by abandoning the use of manual tweak parameters.

In this way, we will provide full automation and reproducibility: from fluorescence movie to trajectories endowed with mis-linking probabilities and estimated model parameters.

To achieve this aim, we will use biophysics modeling and Bayesian analysis for jointly linking detected positions of biomolecules (dots) between time frames AND estimating parameters. We will also decrease computational times for an existing photo-physics-based Bayesian method for dot location.

We will apply our method on tracking data of biomolecules in nanofluidic devices and on experimental data of T-cell receptor dynamics, the latter an important process in immune response where the lack of theoretical tools is pressing.

A PhD student will be hired to implement our methods, which will have as main unique feature that it outputs mis-linking probabilities for trajectories.