How to use HREX to correctly calculate the free energy difference.

[AnguseZhang]

Dear Plumed team, I tried to use HREX to perform FEP-REST2. If we choose a heat region (using plumed partial_tempering), the topologies between different replicas are different. However, when calculating potential energy difference (needed by BAR or MBAR) , Gromacs seems to only use the local topology to compare replicas with different lambdas. So the results from dhdl.xvg are incorrect. As a result, for months I have to use rerun to perform a correction and it's time-consuming. I tried to fix the code by myself , but the code is too complicated and the progress is little. Can you give me some instructions? Many thanks !

[GiovanniBussi]

Hi,

as far as I know, the only way to obtain the desired factors is using mdrun -rerun. It should not be too expensive though, since you can do it with a skipped trajectory.

Giovanni