[gmx-users] All-bonds vs h-bonds contraints

[gpat at bioacademy.gr]

Hello

Could you please clarify what is the reason for adopting the "all-bonds"
constraints as a standard practice in gmx protein simulations?

I mean, the argument that "the time step in molecular simulations can be
increased by a factor of 4 by replacing the bond vibrations by holonomic
constraints' (LINCS paper) seems valid to me in the case of bonds that
contain hydrogen (10 fs vibration period for a C-H bond). Why there is a
need (if any) to turn ALL bonds to constraints?

Furthermore, has anybody seen a difference in the ps/ns protein dynamics
when switching from all-bonds to h-bonds constraints? My opinion is that
if bond vibrations take place in the fs timescale AND are uncoupled to the
slower degrees of freedom , the choice of constraints shouldn't matter,
apart from the computation cost of course.

I would really appreciate any comments on these issues.

Many thanks in advance.
George Patargias