From: Peter Freddolino (petefred_at_umich.edu)
Date: Mon Nov 30 2020 - 19:28:16 CST
NAMDenergy is going to give you the interaction energy that the force field
gives you; no more, and no less. The pairwise additive force fields that
we're typically talking about here don't have any more complex hydrogen
bonding terms , and in particular don't take hydrogen bonding angles into
account (this is a well known aspect of the approximations made by these
force fields). If you want to, you could instead extract the distances and
angles from the trajectory using a tcl script and either analyze them
directly, or look at them in the context of some more complex functional
form for hydrogen bonding energies.
On Mon, Nov 30, 2020 at 1:02 PM Francesco Pietra <chiendarret_at_gmail.com>
> I am aware of the series of posts by Madhavi-Fiorin-Freddolino about
> NamdEnergy and Nonbonded.
> Related to that, I would like to make known what I am observing with a
> protein and its peptide ligand where, according to X-ray diffraction from
> the crystal at very low temperature, several possible H-bonds may be
> involved in stabilizing the receptor-ligand complex.
> I carried out a series DFT QMMM (PBE0 D3BJ def2-SVP level) in explicit
> water at 310K for the atoms presumably involved in these H-bonds. All
> crystal water was retained in MD and included in QMMM when involved. Then,
> the QMMM-only psf/dcd was analyzed by implicit solvent NamdEnergy with
> default cutoff and dielectric according to the values suggested by Li et al
> JCTC 2013 for the residues involved (around 20).
> I find the resulting Nonbonded Energy values (highest -3.5, lowest -0.3
> kcal/mol) mostly unrelated to what is expected from H-bonds on visualizing
> the trajectory (I mean H-O or H-N distances and related angles). My
> question is whether there is a way to tune the NamdEnergy plugin better in
> order to emphasize the role of H-bonding.
> Thanks for advice
> francesco pietra
This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:10 CST