Reaction Boost (targeted MD): RMSD¶
Important
This tutorial requires AMS2024 or later.
The reaction boost (targeted MD) feature can be used to gradually transform one system into another during molecular dynamics.
This feature can be used, for example, to
sample an interesting range of structures for training machine learning potentials,
get reasonable initial guesses for transition states
In this tutorial we will set the reaction boost method to RMSD
, which means
that the method will transform the cartesian coordinates of one system into the target system.
This can push the simulation over the (possibly large) reaction barriers.
For more information about reaction boost (targeted MD) and how to set up these calculations, see Reaction Boost (targeted MD): bond-making and breaking.
Cu atom diffusion on/in Ag(110)¶
ReactionBoostSurfaceRMSD.run
The example loads the ASE (Atomic Simulation Environment) engine with the EMT calculator.
ASE is a Python library that is included with AMS.
AT base pair dissociation in water¶
ReactionBoostATDissociation.run
The example loads the ForceField engine with the GAFF force force field, including automatic atom typing.
On the Model → MD… → Reaction Boost panel, you can see that the reaction boost is only applied to the AT region, which contains the the two AT molecules.
When you switch the view between the two input structures, you can see that the water molecules are in completely different positions. However, no restraints are applied to the water molecules, so it doesn’t matter where they are located in the target structure.
You can also see that Pre equilibration steps is set to 5000
. This
means that the first 5000 steps of the MD simulation are run without any
restraints. The boosting only starts at step 5000.