Modeling phase separation of biomolecular condensates with data-driven mass-conserving reaction-diffusion systems

Phase separation, as one important type of pattern formation, plays a critical role in regulating cellular processes and sustaining ecological resilience. Mass-conserving reaction-diffusion (MCRD) models have been proposed to capture the underlying principles of phase separation. However, previous studies have largely established only phenomenological analogies between MCRD dynamics and phase separation. Here, we identify an experimental model system based on the double-stranded DNA-human protein p53 interactive co-condensate (DPIC). Importantly, all parameters required for the MCRD model can be independently and directly measured in this system, without reliance on parameter estimation or unverified assumptions. We demonstrate that (1) the DPICs serve as an ideal experimental system for establishing a direct and quantitative bridge between experimental DPICs and the MCRD framework and (2) the MCRD model captures more than just a phenomenological resemblance to phase separation, and quantitatively reproduces the underlying dynamics.