Hi,

We were carrying out in vacuo energy minimization studies of a protein dimer (which is experimentally proven to be a dimer). Earlier, the same work has been done in our lab using an older version of GROMACS (4.5.5) and used Group cutoff schemes with coulomb type= cutoff and with no pbc.

When we reinitiated the work again and have to use the Gromacs 5.0.4, the default cutoff scheme is changed to Verlet. We are observing that using Verlet cutoff scheme, the monomers dissociate from each other which is not the case even in this version when using Group cutoff scheme.

I searched for literatures and found out the differences are probably in the pairlist generation. In my graduate courses, I have read about energy drift in molecular dynamics simulation and is aware (though not in details) that Verlet algorithm has something to do with it.

Can anyone elucidate on this problem? The minimization runs fine and the protein remains dimerized when using Group cutoff. This happens even after solvation. We have used an xyz pbc and grid neighbour searching type with default fourier spacing and rlist as we have not mentioned the last two parameters explicitly in the mdp file.

I want to know the theory in play behind this. Please help.

More Souparno Adhikary's questions See All
Similar questions and discussions