[gmx-users] Symmetry corrections in FEP

[Dries Van Rompaey]

Dear gmx-users,

I'm analysing set of FEP simulations where a water molecule is decoupled from a protein. I’m using the orientational restraints as suggested by Boresch et al (doi 10.1021/jp0217839). In my simulations, the water molecule interconverts between two equivalent orientations at the weakly constrained lambda stages, but as the constraints get stronger, the equivalent orientation is no longer sampled. It’s unclear to me how the symmetry correction should be handled in this case. In the literature I have only found Mobley et al’s 2006 paper (doi 10.1063/1.2221683), which mentions that symmetry corrections may need to be applied to each pair of neighbouring simulations when equivalent orientations are only sampled in some of the simulations. In most papers where water molecules are decoupled (for instance Lu et al, doi 10.1021/ja058042g) it seems a complete symmetry correction is applied to all water molecules, regardless of sampling.

I’d be very grateful if someone could shed some light on this issue. 
Thanks in advance,

Dries