Speaker
Description
Molecular dynamics (MD) simulations have been proven essential for elucidating the hierarchical interplay of molecular interaction patterns on different time and length scales. To study the material properties of biomolecular condensates and ultimately their biological function, simulations of large systems require a vast amount of computational resources. Gaining insight into minute details such as the transient dynamics of IDPs requires excessive contact analysis on the atomistic level - a present bottleneck in both storage and processing of large trajectory datasets. Meeting the growing demand for machine-learning analysis for revealing complex contact patterns and property prediction such datasets must be consistent, accessible, and reproducible. We present a cascade computing framework designed as a FAIR-compliant workflow to leverage full pairwise contact information from MD trajectories. Designed for high-performance computing (HPC) clusters, it facilitates the calculation of pairwise contact data through a high degree of parallelization into a contact record that feeds individual consecutive downstream tasks. We apply this framework to investigate the molecular drivers of liquid–liquid phase separation (LLPS) in the foci-forming region (FFR) of MUT-16. Analyzing an aggregate of 10 μs of atomistic simulations—generated via backmapping from Martini 3 coarse-grained models—we utilized the cascade computing package to systematically quantify residue-resolved contact frequencies alongside interaction persistence times. This dual approach enabled a direct comparison between the prevalence of interactions and their lifetimes. We specifically characterized the relative contributions of hydrogen bonding, π–π stacking, cation–π interactions, and salt bridges. Our results reveal that condensate dynamics are governed by specific cation–π interactions and salt bridges rather than non-specific forces. Additionally, the analysis highlighted the critical role of the solvent environment, demonstrating how Na⁺ ions and bulk water modulate stability through ion-mediated bridging. These findings provide deep atomistic insight into the physicochemical principles of MUT-16 condensates and validate the cascade framework as a robust tool for the high-throughput analysis of large-scale protein simulations.