GROMACS version: 2020.6
GROMACS modification: No
I am trying to perform free energy perturbation on a single functional group of a molecule. I have seen free energy can be calculated with the slow-growth method by a lambda value in the MDP options, or by manually changing the topology parameters for each step between “A” and “B” end-states.
For the former method, this would require creating a separate molecule topology description and creating a covalent bond linking this “molecule” to the remainder of the original molecule. This seems like a herculean task for such a simple requirement. Am I thinking about this the right way? Or is there a way to apply lambda to only a single residue of the molecule?
For the latter, I believe this involves changing the values of the [ atomtypes ] section epsilon and [ atoms ] section charge values for the atoms in question, such that the potential equation would look like:
U(\lambda)=4\sqrt{\epsilon_i\epsilon_j(1-\lambda)^2}[(\frac{\sigma_{ij}}{r_{ij}})^{12}-(\frac{\sigma_{ij}}{r_{ij}})^{6}]+\frac{(1-\lambda)q_iq_j}{r_{ij}}
The squared (1-\lambda) in the square-root is to escape the combination rule of \epsilon_{ij}. In short, the (1-\lambda) would be applied to \epsilon and q for each atom belonging to the functional group in question.
I would appreciate some feedback or guidance on this.