The ugly, bad, and good stories of large-scale biomolecular simulations

Molecular modeling of large biomolecular assemblies exemplifies a disruptive area holding both promises and contentions. Propelled by peta and exascale computing, several simulation methodologies have now matured into user-friendly tools that are successfully employed for modeling viruses, membranous nano-constructs, and key pieces of the genetic machinery. We present three unifying biophysical themes that emanate from some of the most recent multi-million atom simulation endeavors. Despite connecting molecular changes with phenotypic outcomes, the quality measures of these simulations remain questionable. We discuss the existing and upcoming strategies for constructing representative ensembles of large systems, how new computing technologies will boost this area, and make a point that integrative modeling guided by experimental data is the future of biomolecular computations. •Growth of computing resources from peta to exascale has bolstered all-atom and Coarse-Grained MD simulations to reach cellular scales.•Large system simulations offer enhanced sampling of the confined subsystems.•Multiscale models uncover a detailed view of diffusion in crowded protein and membrane environments.•Integrative modeling is the future of largescale molecular simulations.