Hi @jesperswillem, we actually don't want shake_solute in place for production FEP using either 1fs or 2fs timesteps. The reason for it is that this setting applies SHAKE to all protein atoms, including, for example C-C, C-N and C-O bonds, which shouldn't be restrained in first place. You can see this setting is turned off on an example file:

With respect to the hydrogens, we're currently making use of the shake_hydrogens as well as shake_solvent. Considering we're typically using TIP3, a rigid water model, it makes sense to have shake_solvent always set to on, regardless of the timestep.

The option shake_hydrogens is left as "STEPTOGGLE" which will be set by QligFEP as "off" for 1fs simulations and "on" for 2fs simulations. Here's a screenshot of the description of the different shake configurations in Q's manual:

Hi @jesperswillem I'm merging this into this new dev branch which I'll use to create the tests