Membraneless organelles, also known as biomolecular condensates, formed via liquid-liquid phase separation (LLPS), have been proposed to play essential roles in diverse cellular processes. Their dysregulation has been implicated in various neurodegenerative diseases, highlighting the need to understand the principles governing their formation. A key challenge is to decode the sequence-encoded rules that tune the thermodynamics and dynamics of biomolecular condensation. Alongside experimental advances, computational modeling at mesoscopic, coarse-grained, and atomistic resolutions has emerged as a powerful approach to probe LLPS. In this review, we summarize recent progress in the predictive modeling of biomolecular phase separation, with a focus on residue-level coarse-grained models that serve as a bridge between mesoscopic models used in field-theoretic simulations and atomistic models. We highlight the approaches adopted in developing models to study LLPS and provide a perspective on directions for future improvement. We conclude by proposing a parameterization strategy that combines multiscale simulations with experimental approaches to uncover the molecular mechanisms underlying condensate formation, maturation, and dysfunction.
Rizuan et al. (2026) studied this question.