Abstract
Computational simulations of biomolecules provide a wealth of information about the thermodynamic landscape of biologically important systems, kinetics of important cellular processes, and the biophysical basis of life. Despite the ubiquity of molecular simulations in biophysical literature, major challenges persist for new practitioners entering the field, and even for experienced computational scientists, in maintaining and distributing their simulation outcomes. Here, we summarize critical obstacles encountered when performing biomolecular simulations and provide best practices for performing simulations that are robust, reproducible, and hypothesis-driven. We also discuss practices that promote improved reproducibility and accessibility using reliable tools and databases.