search

graeter-group / kimmdy

Reactive MD pipeline for GROMACS using Kinetic Monte Carlo / Molecular Dynamics (KIMMDY)
https://graeter-group.github.io/kimmdy/
GNU General Public License v3.0
8 stars 1 forks source link

Plumed file changes shouldnt be hardcoded #442

Open ehhartmann opened 2 months ago

ehhartmann commented 2 months ago

At the moment we automatically remove plumed variables from the plumed file if a bond is broken. This is the behavior we want if we only do homolysis but for homolysis/recombination simulations or something else, we might not want to remove the plumed variables.

https://github.com/hits-mbm-dev/kimmdy/blob/cf625b2239838f40ec3dcc6b86240f57a8258957/src/kimmdy/runmanager.py#L764-L768

Not calculating homolysis rates for broken bonds can also be solved in the homolysis plugin and is a config option at the moment.

            if self.config.check_bound:
                if not atomnrs[1] in top.atoms[atomnrs[0]].bound_to_nrs:
                    continue

What do you think about removing the break_bond_plumed call?

jmbuhr commented 2 months ago

Good point. We just have to check if removing this can lead to us trying to break already broken bonds.

KRiedmiller commented 1 month ago

There are two ways to go about this, having a maximal plumed file with everything ever necessary and never deleting from it, or one in sync with the topology. When we scale to many reactions, option 1 might become unfeasible, also it can slow down the expensive part, the MD. 2 is more complex to implement.