This goes back to the problem with huge differences due to electrostatics. When I asked NAMD to explain how such small differences can create these issues, I was told to make sure that the "net charge" in the mutated area are the same. This is interesting. We ensure that our molecules are net-neutral, but not that "the mutated areas have the same net charge".
Now considering that our alchemical calculation only takes into account the mutated areas, then that somehow makes sense, and is interesting issue. Because do we actually ensure that the mutated region in A and B have the same net charge? ie by taking the average of the superimposed component, we might be shifting it closer to A than to B. I'll add some examples.
This goes back to the problem with huge differences due to electrostatics. When I asked NAMD to explain how such small differences can create these issues, I was told to make sure that the "net charge" in the mutated area are the same. This is interesting. We ensure that our molecules are net-neutral, but not that "the mutated areas have the same net charge".
Now considering that our alchemical calculation only takes into account the mutated areas, then that somehow makes sense, and is interesting issue. Because do we actually ensure that the mutated region in A and B have the same net charge? ie by taking the average of the superimposed component, we might be shifting it closer to A than to B. I'll add some examples.