A single-particle tracking framework for quantifying intracellular partition free energies

Biomolecular condensates are membrane-less organelles that locally enrich specific molecules at spatial sites. A central challenge is determining the physical parameters of individual condensates, as their dimensions are typically below the diffraction limit of conventional microscopy. Single-particle tracking overcomes this limitation, but robust methods for extracting physical observables from trajectory collections remain scarce. Here, we present a framework for inferring the potential difference between molecules inside a condensate and those in the surrounding environment. We introduce a correction that accounts for condensate motion during acquisition, enabling reliable density estimation even from sparse data. After testing on in silico data, we apply our approach to 5 experimental datasets of proteins active in DNA repair and separate them based on the liquid nature of the focus.