Project

General

Profile

Feature #3283

Support for the string method with swarms of trajectories in GROMACS

Added by Oliver Fleetwood about 2 months ago. Updated 26 days ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
Difficulty:
uncategorized
Close

Description

As discussed with the GROMACS developers, we should make sure that we can easily implement enhanced sampling/ensemble methods on top of GROMACS. This feature request specifically outlines what is needed to run the string method with swarms of trajectories. The method will work as a showcase of path sampling and ensemble methods with GROMACS in general.

This feature request is based on the following design document: https://docs.google.com/document/d/10qGoDYoyHSlXPP263egWr8edUe8kvUQURgUdfIqHjfM/edit?usp=sharing, also attached below.

In practice we need to do the following:
  • Make sure the colvars package works with GROMACS for this purpose. We need to be able to restrain a system to some collective variable. Optionally we can use pull code for this.
  • Assert that the python api can start and load trajectories as intended and wait for the simulations to finish.
Nice to haves:
  • Reset the random seed in simulations starting from the same tpr file (currently only supported with the '-multi' option).
  • Don't output the first frame for many replicas starting from the same initial coordinates. In practice this would reduce the total size of xtc files by 50 %.

Subtasks

Feature #3285: Run simulations from the same tpr file with different random seedsNew
Task #3356: Expose setting of random seed to APINew
Feature #3286: Optionally skip initial coordinates from being written to output coordinatesNew

History

#2 Updated by Christian Blau 26 days ago

Just contacted the colvars authores again on what they need within gromacs and pointed to this redmine.

Created https://redmine.gromacs.org/issues/3357 to address colvars specifically.

Also available in: Atom PDF