[gmx-users] How does Gromacs store Free Energy sample data for the BAR/MBAR Method

[Braden Kelly]

Hello,

Gromacs produces an xvg file for each lambda window of a free energy simulation. When using BAR or MBAR are the results in the file actually dH/dL where H is hamiltonian (potential energy) and L is lambda?

This would make sense for Thermodynamic Integration where a derivative is required.

I ask because BAR and MBAR uses the weighted ratio of potential energy differences, not the derivative itself.

Does gromacs version of BAR/MBAR store samples of dH/dL and then calculate the difference in energy between to states as being

U_B = U_A + dH/dL * (L_B - L_A)

which can be written U_B - U_A = dU_A = dH/dL * (L_B - L_A)? (ditto for State B going to A)

where dU_A is the perturbed energy difference between state A and B and state A is the reference state.

Or does Gromacs calculate at each sampling point U_A and U_B to get the difference in potential energy between the two states (dU_A) and it is actually storing the difference in these two values in the xvg file?

Thanks,


Braden


"I feel more like I do now than I did when I first got here."



Braden Kelly

PhD. Candidate, E.I.T

University of Guelph

Biophysics Interdisciplinary Group (BIG)

"I feel more like I do now than I did when I first got here."


[1495766834308_mini_wolf_noir.jpg]