MolSSI Workshop: Molecular Dynamics Software Interoperability

Organizers: Matthieu Chavent​ (IPBS, Toulouse, France), ​John Chodera​ (MSKCC, New York, USA), Karmen Čondić-Jurkić (MSKCC, New York, USA), ​Erik Lindahl​ (KTH, Stockholm, Sweden), ​Jean-Phillip Piquemal (Sorbonne Université, Paris, France)

Location: Brooklyn, New York

Dates: 3-5 November 2019

There are now multiple well-established packages to perform molecular dynamics (MD) simulations (AMBER, CHARMM, GROMACS, NAMD, OPENMM, TINKER, etc.), each with strong but poorly interoperable communities surrounding them. These programs were optimized to work with dedicated sets of inputs/outputs specific to each community. Furthermore, while these programs provide in general a set of basic tools to analyze their own outputs, these tools work mainly with data provided by the related MD program – even if we can notice few exceptions like VMD, cpptraj or MDAnalysis. Thus, it can be often difficult to move systems and data between them. These issues may limit collaborations between communities and force researchers to duplicate effort by implementing and maintaining equivalent tools independently. This also creates barriers to training new scientists, forcing practitioners to relearn everything from scratch when migrating from one MD package to another. Last but not least, transferring data (especially molecular system descriptions) from one program to the other faces significant challenges.

Given the need for modern researchers to produce open, reproducible scientific output, it is now urgent to develop new standards to tackle these long standing issues and break down the walls between MD communities. A recent Bioexcel workshop (Stockholm, Nov 2018) initiated the process of identifying the major problems and potential solutions (see ​https://bioexcel.eu/sdms18-recordings/​). While initially focused on European researchers using the GROMACS software, we now aim to broaden this workshop to all major MD packages and communities and gather representatives of the different MD programs as well as developers of analysis tools. This will help to engage constructive discussions to implement common solutions and trigger collective actions to share analysis tools and data from MD simulations.  This workshop will identify common ground to develop future solutions to easily standardize, share, and analyse data from major MD programs and engage their respective communities around the same objectives.

Our specific objectives are to:

  • develop common solutions to share data in between different MD software;
  • design standard file formats to seamlessly pass from one MD package to the other;
  • identify common tools for analysis and visualization of MD simulations; and
  • establish a community-driven process to ensure development and maintaining of common tools and MD software.

While participation in this workshop was by invitation only, recordings of the workshop’s presentations can be found at this link.

A report on the workshop’s outcomes may be found here.