Phase separation of tunable biomolecular condensates predicted by an interacting particle model

Phase separation is emerging as key principle in the spatiotemporal organization of living cells. Given its relevance in the regulation of numerous biological functions, including gene transcription and chromatin architecture, modeling biomolecular condensation is gaining interest. Yet, most models developed so far rely on specific descriptions and/or experimentally inaccessible properties. Here we propose a theoretical model, where phase separation is explained by means of interaction probabilities between particles. With minimum model requirements, particle condensates emerge above a critical interaction probability. We tested the model predictions with single molecule experiments of tunable transcription factor condensates in the nucleus of living cells. Phase separation, condensate sizes, diffusion behavior, and mobility parameters, quantified by data analysis and machine learning, are fully recapitulated by our model. Our combined theoretical and experimental approach provides a general framework to investigate the biophysical parameters controlling phase separation in living cells and in other soft matter-based interacting systems.