Relationship between couple-moltype and couple-intramol

GROMACS version: 2019.3
GROMACS modification: No

Dear users,
It seems couple-moltype and couple-intramol options are closely related in free energy calculations setting in gromacs.
If couple-moltype is not set and couple-intramol = no, could the system be affected by this setting or not?
Is there a default value for couple-moltype if couple-moltype is not set?

Formally, there is no relationship between the two. One always needs to specify a [moleculetype] to transform, but whether or not its intramolecular interactions are transformed with λ depends on what you’re trying to calculate.

First you should take a look at what couple-intramol is doing. Every setting affects outcomes, and it is important to know how.

No. You need to tell grompp which topology to transform.

Thank you for your explanation. Could I ask two more questions? Thanks!
(1) If we mutate a residue in a protein molecule and do alchemical FEP, we may not need this “couple-moltype”. Because in topology file a residue can not be seen as “a molecule”.
If we do not need “couple-moltype”, the “couple-intramol” seems useless, because we should define molecule first, then Gromacs can know “couple-intramol” in which molecule.
(2) In above alchemical FEP “couple-moltype” absent case, would Gromacs ignore all these three options “couple-intramol=no (default)”; “couple-intramol=yes”; “couple-intramol absent” ?

I have never done a calculation in which couple-moltype is not specified. Even for mutating a protein in which only one residue changes, that is the [moleculetype] that is being transformed; an explicit B-state is specified for the residue that is transformed and the remainder remains in the A-state.

couple-intramol is integral to the calculation, it specifies whether intramolecular force field terms are coupled as a function of λ. For an absolute free energy calculation, it needs to be set to no but for a transformation like you’re describing, in which the molecular structure changes, it should be set to yes.

Thank you. But if we set both ‘couple-moltype=Protein_chain_A’ AND different A,B state in one atom or residue in Protein_chain_A, an error would show up like:

Atom 100 in molecule type ‘Protein_chain_A’ has different A and B state
charges and/or atom types set in the topology file as well as through the
mdp option ‘couple-moltype’. You can not use both these methods

This error suggests “change topol file” method and “couple-moltype” method may not be mixed. They seem to be paralleling method.
Gromacs Document’s “Special topics/Free Energy Implementation” part also indicates it.