I understand that the change of PMF can be calculated via Jarzynski equality if I have a number of replicate steered MD runs for that process. For example, I can pull an inhibitor from its binding site on an enzyme multiple times, and then calculate the change of PMF with Jarzynski equality.
But I wonder if these "replicate SMD runs" should start from the same "initial state", including both atom coordinates and velocities. In my case, I varied the random seed for velocity generation at the beginning (a short NPT equilibration) of each "replicate SMD run", so that I can have 12 different "replicate SMD runs". The consequences of this are very different initial velocities and slightly different initial coordinates for each replicate SMD run. So can I still use Jarzynski equality to calculate deltaPMF?
If the same "initial state" is necessary, then how could I have "replicate SMD runs" that are NOT identical to each other?
In addition, the "final states" in each "replicate SMD runs" are of course different. Does this NOt matter? Or is it that all those different "initial states" are considered to belong to one ensemble, "inhibitor bound", whereas all these different "final states" are considered to belong to the other ensemble, "inhibitor unbound"? So the deltaPMF obtained would be the difference between "inhibitor bound" and "inhibitor unbound"?
- Badges
- Science topic