GROMACS version: 2019
GROMACS modification: No
I have been running a simulation with free-energy coupling, but realised that for lambda=1, which in my case is the end state with interactions fully turned on, the results where different compared to a regular MD run without the free-energy coupling (see attached ion-water RDFs).
These are rather long MD runs (~30ns), and even though the RDFs are not smooth due to the fact that there is only one ion pair, the shape is converged and is not changing when running even longer. Is there anyone that has stumbled upon something similar/any idea of what could be causing the difference?
Best,
Vidar
You did not state what you are decoupling and between what groups you computed the RDF. But in general interactions are what determine how the RDF looks like, so isn’t this exactly expected behavior?
Thank you for the reply. In my case I simultaneously create a potassium ion and a thiocyanate ion in explicit water. The attached RDF is for the end state, since I am only interested in the fully coupled state. The RDF is between the carbon in thiocyanate and the oxygen in water. Free energy coupling at lambda=1 and regular MD for the same state should give the same result, but in my case they do not.
Best regards,
Vidar
It turned out to be a mistake in the input file (couldn’t find out how to close the discussion so posting instead). The bonded potentials in the force field I had copied had a lambda b state defined with an angle not consistent with the regular MD.
Best,
Vidar