How reproducible could be the results from a MD in Gromacs

User discussions
1

Dear Gromacs users,

I have a general question about Molecular Dynamics and the reproducibility of its results.

I am running Simulated Annealing-Molecular Dynamics and in the .mdp I turn off the generation of velocities.

I know that usually MD generate random seeds for the initial velocities so in theory if you repeat the simulation highly probably the results are not going to be the same.

But what about the behavior of the molecular properties?: for example the coefficient of variation of dihedral angles, the folding of a protein?.

Kindly,

Lazaro

2

In general, you should expect that ensemble properties will be the same across independent simulations. Things like folding pathways will never be identical but should generally have the same features (e.g. nucleation of secondary structure elements and order of tertiary contacts).