Single particle tracking with compressive sensing using progressive refinement method on sparse recovery (spt-PRIS)

Abstract

Single particle tracking (SPT) is an indispensable tool for scientific studies. However, SPT for datasets with a high density of particles is still challenging, especially for the study of particle interactions where the point spread functions (PSFs) are overlapping. In this study, we present spt-PRIS, a new SPT solution where we apply compressive sensing to SPT by integrating the progressive refinement method on sparse recovery (PRIS) into the framework of the state-of-the-art SPT algorithm (uTrack). We systematically characterized and validated spt-PRIS performance using simulations, applied it to the experimental data of membrane-bound KRAS4b proteins in either 2-lipid or 8-lipid membrane supported lipid bilayers (SLB), and compared the results to the conventional method (uTrack). Our results show that spt-PRIS is effective for SPT when the data contains overlapping PSFs and provides unprecedented information about KRAS4b subpopulations. spt-PRIS is helpful for a broad range of scientific studies where precise and fast high-density localization is beneficial. spt-PRIS is also flexible for extensions for multi-species, multi-multi-channel, and multi-dimensional SPT methods with the generalization of PRIS reconstruction schemes.