The martini 3 ecosystem for coarse-grained simulations.

Abstract

Coarse-grained molecular dynamics (MD) simulations using the Martini force-field have become a powerful tool in studying biomolecular systems at near to all-atom resolution. About two years ago we put forward a complete reparameterization of the original Martini force-field. The Martini 3 force-field comprises rebalanced non-bonded interaction parameters as well as an improved parametrization strategy for generating molecular parameters. So far Martini 3 has been successfully applied in numerous studies ranging from protein ligand binding for computer-aided drug design to simulations of large-scale systems such as the SARS-Cov2 viral envelope. As exemplified by these studies, the trend goes towards high-throughput sampling studies as well as large-scale biomolecular systems with the set goal of simulating whole cells. To prepare the Martini universe for the next simulation era, advancement in three aspects is of fundamental importance: 1) development of automated workflows that enable fast and efficient setup of complex systems; 2) extension and improvement of available molecule and force-field parameters; 3) management of feedback from the research community. To this end we created the vermouth python library as a unified framework for both program development and parameter management. Based on vermouth, tools such as martinize2, polyply, or fast-forward have been created, which facilitate setting up Martini simulations. In addition, improved parameters for small molecules and carbohydrates have been released with updates for other classes of biomolecules such as lipids underway. Lastly, by openly hosting all programs and parameters on GitHub, we give the research community an opportinuty to provide feedback and contribute to developments. Here we present how these individual developments make up the Martini 3 ecosystem for coarse-grained simulations and our current goals for future advances.